Supplemental Material

1 Prior ESM studies on affective reactivity

2 Data preprocessing

Data preprocessing was conducted largely in accordance with Schoevers et al. (2020). Several checks were performed on the EMA data before using these in further statistical analyses. First, we checked whether responses exceeded the minimum or maximum score (<1 or >7) (0 observations, 0 participants). Second, exceedance of the maximum response time was checked. Respondents were instructed to fill-out the questionnaires as soon as possible after receiving the text message (beep), preferably within 15 minutes but no later than 60 minutes. The patient received a reminder after 30 minutes. We coded observations as missing if they were not uploaded within 65 minutes after the participants were invited by a text message to fill-out the questionnaire (18 observations, from 17 participants). Third, we identified observations that are missing because individuals opened the questionnaire but did not fill it in (0 observations, 0 participants) or because of technical failure (e.g., server down n=19). Fourth, participants with a response rate below 50% (35 of 70 observations) were excluded from the analyses (n = 8). Fifth, participants were excluded if they showed no variation on the positive affect or negative affect scales (1 participant).

3 Data preparation in R

rm(list=ls(all=T))
library(plyr)
library(ggplot2)
library(moments)
library(knitr)
library(dplyr)
library(ggpubr)
library(FSA)
library(lme4)
library(lmerTest)
library(reshape2)
library(stringr)
library(car)
library(effects)
library(DHARMa)
library(GLMMadaptive)

load("/Users/linovonklipstein/Documents/Research Data/NESDA/after preprocessing/formatted data.Rdata")

Variables in loaded dataset (emaD):
pident = participant identification number
negA = negative affect (range 1-7)
posA = positive affect (range 1-7)
posE = positive event (dichotomous; 0-“no”, 1-“yes”)
negE = negative event (dichotomous; 0-“no”, 1-“yes”)
negA_lag1 = NA at t-1 (set to missing at first measurement of a day)
posA_lag1 = PA at t-1 (set to missing at first measurement of a day)

# reorder group factor for clarity
emaD$depgroup <- factor(emaD$depgroup, levels = c("control", "remitted", "current"))

# create second group factor, where "current" is the first group (and will serve as reference group in some analyses)
emaD$depgroup2 <- factor(emaD$depgroup, levels = c("current", "remitted", "control"))

# calculate person-mean centered event variables
emaD <- emaD %>%
  group_by(pident) %>%
  mutate(negE.pm = mean(negE, na.rm = T),
         negE.pmc = negE - negE.pm,
         posE.pm = mean(posE, na.rm = T),
         posE.pmc = posE - posE.pm) %>%
  ungroup()

# calculate person-mean centered lagged affect variables
emaD <- emaD %>%
  group_by(pident) %>%
  mutate(negA_lag1.pm = mean(negA_lag1, na.rm = T),
         negA_lag1.pmc = negA_lag1 - negA_lag1.pm,
         posA_lag1.pm = mean(posA_lag1, na.rm = T),
         posA_lag1.pmc = posA_lag1 - posA_lag1.pm) %>%
  ungroup()

# substract 1 from affect variables to make minimum value 0 (necessary for two-part models, but used in all models for consistency)
emaD$negA <- emaD$negA-1
emaD$posA <- emaD$posA-1

4 Analysis step 1: Tests of group differences in sample characteristics

4.1 Age

# aov_age <- aov(fage ~ depgroup, data = wavesD)
# summary(aov_age)
# plot(aov_age, 1)
# plot(aov_age, 2) 
# leveneTest(fage ~ depgroup, data = wavesD) # residuals not normal -> kruskal-wallis
kruskal.test(fage ~ depgroup, data = wavesD)

## 
##  Kruskal-Wallis rank sum test
## 
## data:  fage by depgroup
## Kruskal-Wallis chi-squared = 5.3245, df = 2, p-value = 0.06979

4.2 Gender

gender_table <- ddply(wavesD, "depgroup", summarise,
      female_N = sum(sex == "female"),
      male_N = sum(sex == "male")
      )
chisq.test(gender_table[,2:3])

## 
##  Pearson's Chi-squared test
## 
## data:  gender_table[, 2:3]
## X-squared = 1.1503, df = 2, p-value = 0.5626

4.3 Years of education

aov_eduyears <- aov(fedu ~ depgroup, data = wavesD)
summary(aov_eduyears)

##              Df Sum Sq Mean Sq F value  Pr(>F)   
## depgroup      2     95   47.50   5.181 0.00607 **
## Residuals   343   3145    9.17                   
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

# plot(aov_eduyears, 1)
# plot(aov_eduyears, 2)
# leveneTest(fedu ~ depgroup, data = wavesD) # variance of residuals is homogeneous and they are approximately normal -> stick with ANOVA

4.4 IDS-SR

# aov_ids <- aov(fids ~ depgroup, data = wavesD)
# summary(aov_ids)
# plot(aov_ids, 1)
# plot(aov_ids, 2) 
# leveneTest(fids ~ depgroup, data = wavesD) # residuals not normal -> Kruskal-Wallis
kruskal.test(fids ~ depgroup, data = wavesD)

## 
##  Kruskal-Wallis rank sum test
## 
## data:  fids by depgroup
## Kruskal-Wallis chi-squared = 137.18, df = 2, p-value < 2.2e-16

4.5 BAI

# aov_bai <- aov(fbaiscal ~ depgroup, data = wavesD)
# summary(aov_bai)
# plot(aov_bai, 1)
# plot(aov_bai, 2)
# leveneTest(fbaiscal ~ depgroup, data = wavesD) # residuals not normal -> Kruskal-Wallis
kruskal.test(fbaiscal ~ depgroup, data = wavesD)

## 
##  Kruskal-Wallis rank sum test
## 
## data:  fbaiscal by depgroup
## Kruskal-Wallis chi-squared = 88.087, df = 2, p-value < 2.2e-16

4.6 Number of positive events

posE_byID <- ddply(emaD, "pident", summarise,
      Event = sum(posE, na.rm = T),
      NoEvent = sum(posE == 0, na.rm = T),
      Missing = sum(is.na(posE)),
      Group = depgroup[1])

# aov_posE <- aov(Event ~ Group, data = posE_byID)
# summary(aov_posE)
# plot(aov_posE, 1)
# plot(aov_posE, 2)
# leveneTest(Event ~ Group, data = posE_byID) # residuals not normal -> Kruskal-Wallis
kruskal.test(Event ~ Group, data = posE_byID)

## 
##  Kruskal-Wallis rank sum test
## 
## data:  Event by Group
## Kruskal-Wallis chi-squared = 4.4294, df = 2, p-value = 0.1092

4.7 Number of negative events

negE_byID <- ddply(emaD, "pident", summarise,
      Event = sum(negE, na.rm = T),
      NoEvent = sum(negE == 0, na.rm = T),
      Missing = sum(is.na(negE)),
      Group = depgroup[1])

# aov_negE <- aov(Event ~ Group, data = negE_byID)
# summary(aov_negE)
# plot(aov_negE, 1)
# plot(aov_negE, 2)
# leveneTest(Event ~ Group, data = negE_byID) # residuals not normal -> Kruskal-Wallis
kruskal.test(Event ~ Group, data = negE_byID)

## 
##  Kruskal-Wallis rank sum test
## 
## data:  Event by Group
## Kruskal-Wallis chi-squared = 12.757, df = 2, p-value = 0.001698

5 Analysis step 2: Within-person means, standard deviations, and skewness

Because these within-person statistics were not normally distributed, we used non-parametric tests to statistically compare them between groups. The Kruskal-Wallis test was used as an omnibus test for group differences and Dunn’s test was subsequently used to test for pairwise comparisons. We employed the false discovery rate to correct p-values and protect against false positive inflation.

5.1 Test code and p-value correction

withinp_descriptives <- ddply(emaD, "pident", summarise,
      M_NA = mean(negA, na.rm = T),
      SD_NA = sd(negA, na.rm = T),
      skew_NA = skewness(negA, na.rm = T),
      M_PA = mean(posA, na.rm = T),
      SD_PA = sd(posA, na.rm = T),
      skew_PA = skewness(posA, na.rm = T),
      Group = depgroup[1]
)

kruskal_M_NA <- kruskal.test(M_NA ~ Group, data = withinp_descriptives)
kruskal_M_PA <- kruskal.test(M_PA ~ Group, data = withinp_descriptives)
kruskal_SD_NA <- kruskal.test(SD_NA ~ Group, data = withinp_descriptives)
kruskal_SD_PA <- kruskal.test(SD_PA ~ Group, data = withinp_descriptives)
kruskal_skew_NA <- kruskal.test(skew_NA ~ Group, data = withinp_descriptives)
kruskal_skew_PA <- kruskal.test(skew_PA ~ Group, data = withinp_descriptives)

dunn_M_NA <- dunnTest(M_NA ~ Group, data = withinp_descriptives, method="none")
dunn_M_PA <- dunnTest(M_PA ~ Group, data = withinp_descriptives, method="none")
dunn_SD_NA <- dunnTest(SD_NA ~ Group, data = withinp_descriptives, method="none")
dunn_SD_PA <- dunnTest(SD_PA ~ Group, data = withinp_descriptives, method="none")
dunn_skew_NA <- dunnTest(skew_NA ~ Group, data = withinp_descriptives, method="none")
dunn_skew_PA <- dunnTest(skew_PA ~ Group, data = withinp_descriptives, method="none")


p_values_tests <- data.frame(model = c("kruskal_M_NA",
                                 "kruskal_M_PA",
                                 "kruskal_SD_NA",
                                 "kruskal_SD_PA",
                                 "kruskal_skew_NA",
                                 "kruskal_skew_PA",
                                 rep("dunn_M_NA",3),
                                 rep("dunn_M_PA",3),
                                 rep("dunn_SD_NA",3),
                                 rep("dunn_SD_PA",3),
                                 rep("dunn_skew_NA",3),
                                 rep("dunn_skew_PA",3)), 
                       p = c(kruskal_M_NA$p.value,
                             kruskal_M_PA$p.value,
                             kruskal_SD_NA$p.value,
                             kruskal_SD_PA$p.value,
                             kruskal_skew_NA$p.value,
                             kruskal_skew_PA$p.value,
                             dunn_M_NA$res$P.unadj,
                             dunn_M_PA$res$P.unadj,
                             dunn_SD_NA$res$P.unadj,
                             dunn_SD_PA$res$P.unadj,
                             dunn_skew_NA$res$P.unadj,
                             dunn_skew_PA$res$P.unadj)
                       )
p_values_tests$p_adj <- p.adjust(p_values_tests$p, method = "BH")

5.2 Kruskal-Wallis omnibus tests

5.2.1 Within-person means

5.2.1.1 NA

## 
##  Kruskal-Wallis rank sum test
## 
## data:  M_NA by Group
## Kruskal-Wallis chi-squared = 102.49, df = 2, p-value < 2.2e-16

5.2.1.2 PA

## 
##  Kruskal-Wallis rank sum test
## 
## data:  M_PA by Group
## Kruskal-Wallis chi-squared = 96.192, df = 2, p-value < 2.2e-16

5.2.2 Within-person standard deviations

5.2.2.1 NA

## 
##  Kruskal-Wallis rank sum test
## 
## data:  SD_NA by Group
## Kruskal-Wallis chi-squared = 99.825, df = 2, p-value < 2.2e-16

5.2.2.2 PA

## 
##  Kruskal-Wallis rank sum test
## 
## data:  SD_PA by Group
## Kruskal-Wallis chi-squared = 43.042, df = 2, p-value = 1.038e-09

5.2.3 Within-person skewness

5.2.3.1 NA

## 
##  Kruskal-Wallis rank sum test
## 
## data:  skew_NA by Group
## Kruskal-Wallis chi-squared = 62.144, df = 2, p-value = 9.612e-14

5.2.3.2 PA

## 
##  Kruskal-Wallis rank sum test
## 
## data:  skew_PA by Group
## Kruskal-Wallis chi-squared = 10.494, df = 2, p-value = 0.006015

5.3 Dunn’s tests for pairwise comparisons

5.3.1 Within-person means

5.3.1.1 NA

##           Comparison          Z        P.adj
## 1  control - current -10.046567 2.282970e-22
## 2 control - remitted  -6.175364 1.320228e-09
## 3 current - remitted   5.915153 5.684002e-09

5.3.1.2 PA

##           Comparison         Z        P.adj
## 1  control - current  9.734479 1.179197e-21
## 2 control - remitted  5.972300 4.318742e-09
## 3 current - remitted -5.740963 1.506238e-08

5.3.2 Within-person standard deviations

5.3.2.1 NA

##           Comparison         Z        P.adj
## 1  control - current -9.720873 1.179197e-21
## 2 control - remitted -6.923434 1.175535e-11
## 3 current - remitted  4.916322 1.244967e-06

5.3.2.2 PA

##           Comparison         Z        P.adj
## 1  control - current -6.226920 1.037879e-09
## 2 control - remitted -4.940614 1.168152e-06
## 3 current - remitted  2.718894 7.145525e-03

5.3.3 Within-person skewness

5.3.3.1 NA

##           Comparison         Z        P.adj
## 1  control - current  7.832951 1.633881e-14
## 2 control - remitted  4.741061 2.834692e-06
## 3 current - remitted -4.674519 3.721809e-06

5.3.3.2 PA

##           Comparison         Z       P.adj
## 1  control - current -3.222490 0.001524982
## 2 control - remitted -1.920830 0.054753168
## 3 current - remitted  1.948341 0.053607837

6 Analysis step 3: Replication analysis

6.1 Model code and p-value correction

# main effect models
Model_NA_negE <- lmer(negA ~ negE.pmc + negA_lag1.pmc + negE.pm + (negE.pmc + negA_lag1.pmc|pident) , data = emaD)
Model_NA_posE <- lmer(negA ~ posE.pmc + negA_lag1.pmc + posE.pm + (posE.pmc + negA_lag1.pmc|pident) , data = emaD)
Model_PA_negE <- lmer(posA ~ negE.pmc + posA_lag1.pmc + negE.pm + (negE.pmc + posA_lag1.pmc|pident) , data = emaD)
Model_PA_posE <- lmer(posA ~ posE.pmc + posA_lag1.pmc + posE.pm + (posE.pmc + posA_lag1.pmc|pident) , data = emaD)

# group difference models
Model_NA_negE_groups <- lmer(negA ~ negE.pmc + negA_lag1.pmc + depgroup + negE.pm + depgroup*negE.pmc + depgroup*negE.pm + (negE.pmc + negA_lag1.pmc|pident) , data = emaD, control=lmerControl(optimizer = "bobyqa", optCtrl=list(maxfun=2e6)))
Model_NA_negE_groups2 <- lmer(negA ~ negE.pmc + negA_lag1.pmc + depgroup2 + negE.pm + depgroup2*negE.pmc + depgroup2*negE.pm + (negE.pmc + negA_lag1.pmc|pident) , data = emaD, control=lmerControl(optimizer = "bobyqa", optCtrl=list(maxfun=2e6)))

Model_NA_posE_groups <- lmer(negA ~ posE.pmc + negA_lag1.pmc + depgroup + posE.pm + depgroup*posE.pmc + depgroup*posE.pm + (posE.pmc + negA_lag1.pmc|pident) , data = emaD)
Model_NA_posE_groups2 <- lmer(negA ~ posE.pmc + negA_lag1.pmc + depgroup2 + posE.pm + depgroup2*posE.pmc + depgroup2*posE.pm + (posE.pmc + negA_lag1.pmc|pident) , data = emaD)

Model_PA_negE_groups <- lmer(posA ~ negE.pmc + posA_lag1.pmc + depgroup + negE.pm + depgroup*negE.pmc + depgroup*negE.pm + (negE.pmc + posA_lag1.pmc|pident) , data = emaD)
Model_PA_negE_groups2 <- lmer(posA ~ negE.pmc + posA_lag1.pmc + depgroup2 + negE.pm + depgroup2*negE.pmc + depgroup2*negE.pm + (negE.pmc + posA_lag1.pmc|pident) , data = emaD)

Model_PA_posE_groups <- lmer(posA ~ posE.pmc + posA_lag1.pmc + depgroup + posE.pm + depgroup*posE.pmc + depgroup*posE.pm + (posE.pmc + posA_lag1.pmc|pident) , data = emaD)
Model_PA_posE_groups2 <- lmer(posA ~ posE.pmc + posA_lag1.pmc + depgroup2 + posE.pm + depgroup2*posE.pmc + depgroup2*posE.pm + (posE.pmc + posA_lag1.pmc|pident) , data = emaD)

# p-value correction using false discovery rate
p_values <- data.frame(model = c(rep("Model_NA_negE", 4),
                                 rep("Model_NA_posE", 4),
                                 rep("Model_PA_negE", 4),
                                 rep("Model_PA_posE", 4),
                                 rep("Model_NA_negE_groups", 10),
                                 rep("Model_NA_negE_groups2", 10),
                                 rep("Model_NA_posE_groups", 10),
                                 rep("Model_NA_posE_groups2", 10),
                                 rep("Model_PA_negE_groups", 10),
                                 rep("Model_PA_negE_groups2", 10),
                                 rep("Model_PA_posE_groups", 10),
                                 rep("Model_PA_posE_groups2", 10)
                                 ),
                       
                       p = c(summary(Model_NA_negE)$coefficients[,5],
                             summary(Model_NA_posE)$coefficients[,5],
                             summary(Model_PA_negE)$coefficients[,5],
                             summary(Model_PA_posE)$coefficients[,5],
                             summary(Model_NA_negE_groups)$coefficients[,5],
                             summary(Model_NA_negE_groups2)$coefficients[,5],
                             summary(Model_NA_posE_groups)$coefficients[,5],
                             summary(Model_NA_posE_groups2)$coefficients[,5],
                             summary(Model_PA_negE_groups)$coefficients[,5],
                             summary(Model_PA_negE_groups2)$coefficients[,5],
                             summary(Model_PA_posE_groups)$coefficients[,5],
                             summary(Model_PA_posE_groups2)$coefficients[,5]
                             )
                       )

p_values$p_adj <- p.adjust(p_values$p, method = "BH")

6.2 Main effect models

Displayed p-values are corrected for false discovery rate

6.2.1 NA-reactivity to negative events

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: 
## negA ~ negE.pmc + negA_lag1.pmc + negE.pm + (negE.pmc + negA_lag1.pmc |  
##     pident)
##    Data: emaD
## 
## REML criterion at convergence: 19941.5
## 
## Scaled residuals: 
##    Min     1Q Median     3Q    Max 
## -6.542 -0.439 -0.096  0.277  7.543 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr     
##  pident   (Intercept)   0.4598   0.678             
##           negE.pmc      0.1228   0.350    0.30     
##           negA_lag1.pmc 0.0279   0.167    0.43 0.23
##  Residual               0.1658   0.407             
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##               Estimate Std. Error       df t value Pr(>|t|)    
## (Intercept)     0.4728     0.0517 371.2078    9.15  < 2e-16 ***
## negE.pmc        0.4034     0.0231 279.1865   17.46  < 2e-16 ***
## negA_lag1.pmc   0.2877     0.0124 325.7226   23.15  < 2e-16 ***
## negE.pm         1.1382     0.2839 345.4822    4.01  0.00014 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) ngE.pmc ngA_1.
## negE.pmc     0.144               
## ngA_lg1.pmc  0.194  0.128        
## negE.pm     -0.706  0.046   0.034

6.2.1.1 BIC

## [1] 20048.5

6.2.2 NA-reactivity to positive events

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: 
## negA ~ posE.pmc + negA_lag1.pmc + posE.pm + (posE.pmc + negA_lag1.pmc |  
##     pident)
##    Data: emaD
## 
## REML criterion at convergence: 21171.6
## 
## Scaled residuals: 
##    Min     1Q Median     3Q    Max 
## -4.894 -0.446 -0.116  0.250  8.882 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr       
##  pident   (Intercept)   0.4650   0.682               
##           posE.pmc      0.0237   0.154    -0.69      
##           negA_lag1.pmc 0.0264   0.162     0.38 -0.32
##  Residual               0.1818   0.426               
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##               Estimate Std. Error       df t value Pr(>|t|)    
## (Intercept)     0.7785     0.0591 400.1066   13.16   <2e-16 ***
## posE.pmc       -0.1494     0.0118 310.3589  -12.68   <2e-16 ***
## negA_lag1.pmc   0.2932     0.0124 328.6077   23.56   <2e-16 ***
## posE.pm        -0.4781     0.1385 370.2152   -3.45    0.001 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) psE.pmc ngA_1.
## posE.pmc    -0.251               
## ngA_lg1.pmc  0.169 -0.137        
## posE.pm     -0.783 -0.065  -0.004

6.2.2.1 BIC

## [1] 21278.61

6.2.3 PA-reactivity to negative events

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: 
## posA ~ negE.pmc + posA_lag1.pmc + negE.pm + (negE.pmc + posA_lag1.pmc |  
##     pident)
##    Data: emaD
## 
## REML criterion at convergence: 33492.3
## 
## Scaled residuals: 
##    Min     1Q Median     3Q    Max 
## -6.229 -0.504  0.073  0.574  4.446 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr       
##  pident   (Intercept)   0.6440   0.802               
##           negE.pmc      0.1592   0.399     0.23      
##           posA_lag1.pmc 0.0184   0.136    -0.26  0.04
##  Residual               0.3776   0.614               
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##               Estimate Std. Error       df t value Pr(>|t|)    
## (Intercept)     4.0076     0.0628 354.1361   63.86  < 2e-16 ***
## negE.pmc       -0.5675     0.0286 274.3861  -19.84  < 2e-16 ***
## posA_lag1.pmc   0.3259     0.0105 344.9856   30.98  < 2e-16 ***
## negE.pm        -1.7115     0.3529 339.2434   -4.85  3.9e-06 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) ngE.pmc psA_1.
## negE.pmc     0.085               
## psA_lg1.pmc -0.113  0.038        
## negE.pm     -0.722  0.047  -0.014

6.2.3.1 BIC

## [1] 33599.36

6.2.4 PA-reactivity to positive events

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: 
## posA ~ posE.pmc + posA_lag1.pmc + posE.pm + (posE.pmc + posA_lag1.pmc |  
##     pident)
##    Data: emaD
## 
## REML criterion at convergence: 33933.7
## 
## Scaled residuals: 
##    Min     1Q Median     3Q    Max 
## -6.091 -0.478  0.094  0.579  3.833 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr       
##  pident   (Intercept)   0.6659   0.816               
##           posE.pmc      0.0537   0.232    -0.51      
##           posA_lag1.pmc 0.0191   0.138    -0.17 -0.08
##  Residual               0.3901   0.625               
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##               Estimate Std. Error       df t value Pr(>|t|)    
## (Intercept)     3.6014     0.0741 371.0162    48.6   <2e-16 ***
## posE.pmc        0.3428     0.0176 313.1215    19.4   <2e-16 ***
## posA_lag1.pmc   0.3126     0.0107 342.8077    29.1   <2e-16 ***
## posE.pm         0.5517     0.1780 355.3781     3.1   0.0033 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) psE.pmc psA_1.
## posE.pmc    -0.174               
## psA_lg1.pmc -0.067 -0.088        
## posE.pm     -0.803 -0.050  -0.004

6.2.4.1 BIC

## [1] 34040.79

6.3 Models estimating group differences

Displayed p-values are corrected for false discovery rate. The control group is the reference group.

6.3.1 NA-reactivity to negative events

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negA ~ negE.pmc + negA_lag1.pmc + depgroup + negE.pm + depgroup *  
##     negE.pmc + depgroup * negE.pm + (negE.pmc + negA_lag1.pmc |      pident)
##    Data: emaD
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 2e+06))
## 
## REML criterion at convergence: 19865
## 
## Scaled residuals: 
##    Min     1Q Median     3Q    Max 
## -6.547 -0.434 -0.093  0.278  7.540 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr     
##  pident   (Intercept)   0.3430   0.586             
##           negE.pmc      0.1153   0.340    0.21     
##           negA_lag1.pmc 0.0272   0.165    0.31 0.15
##  Residual               0.1659   0.407             
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##                           Estimate Std. Error       df t value Pr(>|t|)    
## (Intercept)                 0.2100     0.0851 352.6675    2.47  0.01955 *  
## negE.pmc                    0.2782     0.0449 293.8440    6.19  4.5e-09 ***
## negA_lag1.pmc               0.2979     0.0124 325.5817   23.99  < 2e-16 ***
## depgroupremitted            0.2085     0.1060 352.2634    1.97  0.06484 .  
## depgroupcurrent             1.0716     0.1354 346.7407    7.91  8.8e-14 ***
## negE.pm                     0.6130     0.6378 347.0996    0.96  0.36552    
## negE.pmc:depgroupremitted   0.1420     0.0540 284.4114    2.63  0.01303 *  
## negE.pmc:depgroupcurrent    0.2486     0.0691 279.6886    3.60  0.00067 ***
## depgroupremitted:negE.pm    0.7021     0.7220 346.0659    0.97  0.36552    
## depgroupcurrent:negE.pm    -1.0293     0.8251 343.7937   -1.25  0.24064    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##                (Intr) ngE.pmc ngA_1. dpgrpr dpgrpc negE.pm ngE.pmc:dpgrpr
## negE.pmc        0.077                                                    
## ngA_lg1.pmc     0.041  0.022                                             
## depgrprmttd    -0.800 -0.061   0.032                                     
## depgrpcrrnt    -0.626 -0.047   0.045  0.505                              
## negE.pm        -0.717  0.043   0.013  0.576  0.451                       
## ngE.pmc:dpgrpr -0.063 -0.831   0.016  0.078  0.041 -0.035                
## ngE.pmc:dpgrpc -0.049 -0.650   0.025  0.041  0.089 -0.027   0.541        
## dpgrprmt:E.     0.633 -0.038  -0.006 -0.719 -0.398 -0.883   0.040        
## dpgrpcrr:E.     0.554 -0.033  -0.008 -0.445 -0.704 -0.773   0.027        
##                ngE.pmc:dpgrpc dpgrpr:E.
## negE.pmc                               
## ngA_lg1.pmc                            
## depgrprmttd                            
## depgrpcrrnt                            
## negE.pm                                
## ngE.pmc:dpgrpr                         
## ngE.pmc:dpgrpc                         
## dpgrprmt:E.     0.024                  
## dpgrpcrr:E.     0.028          0.683

6.3.1.1 BIC

## [1] 20030.44

6.3.2 NA-reactivity to positive events

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negA ~ posE.pmc + negA_lag1.pmc + depgroup + posE.pm + depgroup *  
##     posE.pmc + depgroup * posE.pm + (posE.pmc + negA_lag1.pmc |      pident)
##    Data: emaD
## 
## REML criterion at convergence: 21098.2
## 
## Scaled residuals: 
##    Min     1Q Median     3Q    Max 
## -4.884 -0.452 -0.115  0.253  8.878 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr       
##  pident   (Intercept)   0.3482   0.590               
##           posE.pmc      0.0202   0.142    -0.62      
##           negA_lag1.pmc 0.0257   0.160     0.27 -0.24
##  Residual               0.1818   0.426               
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##                           Estimate Std. Error       df t value Pr(>|t|)    
## (Intercept)                 0.3224     0.0971 379.0579    3.32  0.00163 ** 
## posE.pmc                   -0.0650     0.0219 323.1161   -2.97  0.00494 ** 
## negA_lag1.pmc               0.3028     0.0124 329.6284   24.37  < 2e-16 ***
## depgroupremitted            0.3821     0.1217 378.2093    3.14  0.00297 ** 
## depgroupcurrent             1.1785     0.1540 373.8921    7.65  4.2e-13 ***
## posE.pm                    -0.1786     0.2207 367.9882   -0.81  0.44207    
## posE.pmc:depgroupremitted  -0.1040     0.0266 313.8397   -3.91  0.00021 ***
## posE.pmc:depgroupcurrent   -0.1564     0.0350 323.9748   -4.47  2.1e-05 ***
## depgroupremitted:posE.pm   -0.1214     0.2810 365.3701   -0.43  0.67288    
## depgroupcurrent:posE.pm    -0.7547     0.3980 360.8569   -1.90  0.07416 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##                (Intr) psE.pmc ngA_1. dpgrpr dpgrpc posE.pm psE.pmc:dpgrpr
## posE.pmc       -0.208                                                    
## ngA_lg1.pmc     0.041 -0.025                                             
## depgrprmttd    -0.796  0.164   0.017                                     
## depgrpcrrnt    -0.628  0.130   0.028  0.503                              
## posE.pm        -0.786 -0.058  -0.005  0.627  0.495                       
## psE.pmc:dpgrpr  0.169 -0.823  -0.023 -0.201 -0.108  0.048                
## psE.pmc:dpgrpc  0.128 -0.625  -0.023 -0.104 -0.211  0.037   0.516        
## dpgrprmt:E.     0.618  0.046   0.007 -0.797 -0.389 -0.786  -0.060        
## dpgrpcrr:E.     0.436  0.032   0.004 -0.348 -0.781 -0.555  -0.027        
##                psE.pmc:dpgrpc dpgrpr:E.
## posE.pmc                               
## ngA_lg1.pmc                            
## depgrprmttd                            
## depgrpcrrnt                            
## posE.pm                                
## psE.pmc:dpgrpr                         
## psE.pmc:dpgrpc                         
## dpgrprmt:E.    -0.029                  
## dpgrpcrr:E.    -0.053          0.436

6.3.2.1 BIC

## [1] 21263.67

6.3.3 PA-reactivity to negative events

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: posA ~ negE.pmc + posA_lag1.pmc + depgroup + negE.pm + depgroup *  
##     negE.pmc + depgroup * negE.pm + (negE.pmc + posA_lag1.pmc |      pident)
##    Data: emaD
## 
## REML criterion at convergence: 33401.8
## 
## Scaled residuals: 
##    Min     1Q Median     3Q    Max 
## -6.207 -0.504  0.070  0.576  4.454 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr       
##  pident   (Intercept)   0.4699   0.685               
##           negE.pmc      0.1592   0.399     0.18      
##           posA_lag1.pmc 0.0182   0.135    -0.10  0.08
##  Residual               0.3776   0.614               
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##                           Estimate Std. Error       df t value Pr(>|t|)    
## (Intercept)                 4.5159     0.1014 344.2710   44.54  < 2e-16 ***
## negE.pmc                   -0.4997     0.0573 297.5819   -8.72  6.2e-16 ***
## posA_lag1.pmc               0.3284     0.0105 344.9457   31.22  < 2e-16 ***
## depgroupremitted           -0.5421     0.1268 343.8007   -4.28  4.7e-05 ***
## depgroupcurrent            -1.5230     0.1631 343.1374   -9.34  < 2e-16 ***
## negE.pm                    -1.9858     0.7660 340.3578   -2.59    0.014 *  
## negE.pmc:depgroupremitted  -0.0745     0.0687 284.4730   -1.08    0.312    
## negE.pmc:depgroupcurrent   -0.1480     0.0879 280.7951   -1.68    0.111    
## depgroupremitted:negE.pm    0.3983     0.8681 339.8941    0.46    0.660    
## depgroupcurrent:negE.pm     2.3878     0.9963 341.7719    2.40    0.023 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##                (Intr) ngE.pmc psA_1. dpgrpr dpgrpc negE.pm ngE.pmc:dpgrpr
## negE.pmc        0.058                                                    
## psA_lg1.pmc    -0.018  0.027                                             
## depgrprmttd    -0.800 -0.047  -0.004                                     
## depgrpcrrnt    -0.621 -0.037  -0.005  0.497                              
## negE.pm        -0.719  0.046  -0.003  0.575  0.447                       
## ngE.pmc:dpgrpr -0.049 -0.834   0.001  0.061  0.031 -0.038                
## ngE.pmc:dpgrpc -0.038 -0.652   0.001  0.031  0.077 -0.030   0.544        
## dpgrprmt:E.     0.635 -0.041   0.001 -0.720 -0.394 -0.882   0.044        
## dpgrpcrr:E.     0.553 -0.035   0.001 -0.442 -0.706 -0.769   0.030        
##                ngE.pmc:dpgrpc dpgrpr:E.
## negE.pmc                               
## psA_lg1.pmc                            
## depgrprmttd                            
## depgrpcrrnt                            
## negE.pm                                
## ngE.pmc:dpgrpr                         
## ngE.pmc:dpgrpc                         
## dpgrprmt:E.     0.027                  
## dpgrpcrr:E.     0.031          0.678

6.3.3.1 BIC

## [1] 33567.28

6.3.4 PA-reactivity to positive events

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: posA ~ posE.pmc + posA_lag1.pmc + depgroup + posE.pm + depgroup *  
##     posE.pmc + depgroup * posE.pm + (posE.pmc + posA_lag1.pmc |      pident)
##    Data: emaD
## 
## REML criterion at convergence: 33841.9
## 
## Scaled residuals: 
##    Min     1Q Median     3Q    Max 
## -6.064 -0.479  0.091  0.576  3.825 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr       
##  pident   (Intercept)   0.4849   0.696               
##           posE.pmc      0.0503   0.224    -0.44      
##           posA_lag1.pmc 0.0191   0.138    -0.04 -0.14
##  Residual               0.3902   0.625               
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##                           Estimate Std. Error       df t value Pr(>|t|)    
## (Intercept)                 4.2978     0.1185 358.3443   36.26  < 2e-16 ***
## posE.pmc                    0.2416     0.0334 329.1922    7.24  7.8e-12 ***
## posA_lag1.pmc               0.3149     0.0107 342.8788   29.33  < 2e-16 ***
## depgroupremitted           -0.6790     0.1492 358.0740   -4.55  1.5e-05 ***
## depgroupcurrent            -1.5700     0.1895 358.1907   -8.28  6.8e-15 ***
## posE.pm                     0.0971     0.2727 352.5355    0.36  0.72203    
## posE.pmc:depgroupremitted   0.1252     0.0405 318.6305    3.09  0.00341 ** 
## posE.pmc:depgroupcurrent    0.1886     0.0532 326.7822    3.55  0.00077 ***
## depgroupremitted:posE.pm    0.2833     0.3485 351.2659    0.81  0.44207    
## depgroupcurrent:posE.pm     0.9994     0.4965 351.8726    2.01  0.05903 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##                (Intr) psE.pmc psA_1. dpgrpr dpgrpc posE.pm psE.pmc:dpgrpr
## posE.pmc       -0.151                                                    
## psA_lg1.pmc    -0.007 -0.055                                             
## depgrprmttd    -0.795  0.120  -0.001                                     
## depgrpcrrnt    -0.625  0.095  -0.001  0.497                              
## posE.pm        -0.797 -0.042  -0.001  0.633  0.498                       
## psE.pmc:dpgrpr  0.125 -0.822  -0.006 -0.148 -0.078  0.035                
## psE.pmc:dpgrpc  0.095 -0.625  -0.016 -0.076 -0.158  0.027   0.516        
## dpgrprmt:E.     0.623  0.033  -0.001 -0.808 -0.390 -0.783  -0.043        
## dpgrpcrr:E.     0.438  0.023  -0.001 -0.348 -0.791 -0.549  -0.019        
##                psE.pmc:dpgrpc dpgrpr:E.
## posE.pmc                               
## psA_lg1.pmc                            
## depgrprmttd                            
## depgrpcrrnt                            
## posE.pm                                
## psE.pmc:dpgrpr                         
## psE.pmc:dpgrpc                         
## dpgrprmt:E.    -0.021                  
## dpgrpcrr:E.    -0.037          0.430

6.3.4.1 BIC

## [1] 34007.36

6.4 Models estimating group differences with current group as reference group

Displayed p-values are corrected for false discovery rate.

6.4.1 NA-reactivity to negative events

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negA ~ negE.pmc + negA_lag1.pmc + depgroup2 + negE.pm + depgroup2 *  
##     negE.pmc + depgroup2 * negE.pm + (negE.pmc + negA_lag1.pmc |      pident)
##    Data: emaD
## Control: lmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 2e+06))
## 
## REML criterion at convergence: 19865
## 
## Scaled residuals: 
##    Min     1Q Median     3Q    Max 
## -6.547 -0.434 -0.093  0.278  7.540 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr     
##  pident   (Intercept)   0.3430   0.586             
##           negE.pmc      0.1153   0.340    0.21     
##           negA_lag1.pmc 0.0272   0.165    0.31 0.15
##  Residual               0.1659   0.407             
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##                            Estimate Std. Error       df t value Pr(>|t|)    
## (Intercept)                  1.2816     0.1056 340.8637   12.13  < 2e-16 ***
## negE.pmc                     0.5268     0.0525 269.3857   10.03  < 2e-16 ***
## negA_lag1.pmc                0.2979     0.0124 325.5819   23.99  < 2e-16 ***
## depgroup2remitted           -0.8631     0.1228 339.2484   -7.03  2.7e-11 ***
## depgroup2control            -1.0716     0.1354 346.7403   -7.91  8.8e-14 ***
## negE.pm                     -0.4163     0.5235 332.3737   -0.80  0.44565    
## negE.pmc:depgroup2remitted  -0.1067     0.0604 267.6464   -1.76  0.09822 .  
## negE.pmc:depgroup2control   -0.2486     0.0691 279.6884   -3.60  0.00067 ***
## depgroup2remitted:negE.pm    1.7313     0.6234 336.1270    2.78  0.00868 ** 
## depgroup2control:negE.pm     1.0293     0.8251 343.7934    1.25  0.24064    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##                 (Intr) ngE.pmc ngA_1. dpgrp2r dpgrp2c negE.pm ngE.pmc:dpgrp2r
## negE.pmc         0.100                                                       
## ngA_lg1.pmc      0.090  0.051                                                
## dpgrp2rmttd     -0.855 -0.083  -0.022                                        
## dpgrp2cntrl     -0.778 -0.077  -0.045  0.667                                 
## negE.pm         -0.719  0.013   0.002  0.618   0.561                         
## ngE.pmc:dpgrp2r -0.084 -0.867  -0.014  0.092   0.065  -0.012                 
## ngE.pmc:dpgrp2c -0.075 -0.759  -0.025  0.063   0.089  -0.010   0.659         
## dpgrp2rm:E.      0.604 -0.011   0.005 -0.730  -0.471  -0.840   0.019         
## dpgrp2cn:E.      0.457 -0.008   0.008 -0.393  -0.704  -0.634   0.007         
##                 ngE.pmc:dpgrp2c dpgrp2r:E.
## negE.pmc                                  
## ngA_lg1.pmc                               
## dpgrp2rmttd                               
## dpgrp2cntrl                               
## negE.pm                                   
## ngE.pmc:dpgrp2r                           
## ngE.pmc:dpgrp2c                           
## dpgrp2rm:E.      0.008                    
## dpgrp2cn:E.      0.028           0.533

6.4.2 NA-reactivity to positive events

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negA ~ posE.pmc + negA_lag1.pmc + depgroup2 + posE.pm + depgroup2 *  
##     posE.pmc + depgroup2 * posE.pm + (posE.pmc + negA_lag1.pmc |      pident)
##    Data: emaD
## 
## REML criterion at convergence: 21098.2
## 
## Scaled residuals: 
##    Min     1Q Median     3Q    Max 
## -4.884 -0.452 -0.115  0.253  8.878 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr       
##  pident   (Intercept)   0.3482   0.590               
##           posE.pmc      0.0202   0.142    -0.62      
##           negA_lag1.pmc 0.0257   0.160     0.27 -0.24
##  Residual               0.1818   0.426               
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##                            Estimate Std. Error       df t value Pr(>|t|)    
## (Intercept)                  1.5009     0.1198 369.1824   12.53  < 2e-16 ***
## posE.pmc                    -0.2214     0.0273 324.4995   -8.12  2.8e-14 ***
## negA_lag1.pmc                0.3028     0.0124 329.6284   24.37  < 2e-16 ***
## depgroup2remitted           -0.7964     0.1403 368.1889   -5.68  6.1e-08 ***
## depgroup2control            -1.1785     0.1540 373.8921   -7.65  4.2e-13 ***
## posE.pm                     -0.9333     0.3311 353.4582   -2.82   0.0078 ** 
## posE.pmc:depgroup2remitted   0.0524     0.0311 315.8165    1.68   0.1108    
## posE.pmc:depgroup2control    0.1564     0.0350 323.9748    4.47  2.1e-05 ***
## depgroup2remitted:posE.pm    0.6333     0.3740 354.8576    1.69   0.1108    
## depgroup2control:posE.pm     0.7547     0.3980 360.8568    1.90   0.0742 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##                 (Intr) psE.pmc ngA_1. dpgrp2r dpgrp2c posE.pm psE.pmc:dpgrp2r
## posE.pmc        -0.215                                                       
## ngA_lg1.pmc      0.069 -0.050                                                
## dpgrp2rmttd     -0.851  0.181  -0.016                                        
## dpgrp2cntrl     -0.776  0.166  -0.028  0.662                                 
## posE.pm         -0.783 -0.050   0.001  0.668   0.609                         
## psE.pmc:dpgrp2r  0.186 -0.874   0.007 -0.206  -0.144   0.044                 
## psE.pmc:dpgrp2c  0.167 -0.780   0.023 -0.141  -0.211   0.039   0.682         
## dpgrp2rm:E.      0.693  0.044   0.001 -0.791  -0.539  -0.885  -0.053         
## dpgrp2cn:E.      0.651  0.042  -0.004 -0.556  -0.781  -0.832  -0.037         
##                 psE.pmc:dpgrp2c dpgrp2r:E.
## posE.pmc                                  
## ngA_lg1.pmc                               
## dpgrp2rmttd                               
## dpgrp2cntrl                               
## posE.pm                                   
## psE.pmc:dpgrp2r                           
## psE.pmc:dpgrp2c                           
## dpgrp2rm:E.     -0.035                    
## dpgrp2cn:E.     -0.053           0.737

6.4.3 PA-reactivity to negative events

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: posA ~ negE.pmc + posA_lag1.pmc + depgroup2 + negE.pm + depgroup2 *  
##     negE.pmc + depgroup2 * negE.pm + (negE.pmc + posA_lag1.pmc |      pident)
##    Data: emaD
## 
## REML criterion at convergence: 33401.8
## 
## Scaled residuals: 
##    Min     1Q Median     3Q    Max 
## -6.207 -0.504  0.070  0.576  4.454 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr       
##  pident   (Intercept)   0.4699   0.685               
##           negE.pmc      0.1592   0.399     0.18      
##           posA_lag1.pmc 0.0182   0.135    -0.10  0.08
##  Residual               0.3776   0.614               
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##                            Estimate Std. Error       df t value Pr(>|t|)    
## (Intercept)                  2.9929     0.1278 342.4372   23.42  < 2e-16 ***
## negE.pmc                    -0.6476     0.0667 270.2223   -9.71  < 2e-16 ***
## posA_lag1.pmc                0.3284     0.0105 344.9458   31.22  < 2e-16 ***
## depgroup2remitted            0.9809     0.1487 342.5059    6.60  3.6e-10 ***
## depgroup2control             1.5230     0.1631 343.1374    9.34  < 2e-16 ***
## negE.pm                      0.4020     0.6371 342.2618    0.63    0.546    
## negE.pmc:depgroup2remitted   0.0735     0.0767 266.9850    0.96    0.366    
## negE.pmc:depgroup2control    0.1480     0.0879 280.7952    1.68    0.111    
## depgroup2remitted:negE.pm   -1.9895     0.7568 341.6444   -2.63    0.013 *  
## depgroup2control:negE.pm    -2.3878     0.9963 341.7719   -2.40    0.023 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##                 (Intr) ngE.pmc psA_1. dpgrp2r dpgrp2c negE.pm ngE.pmc:dpgrp2r
## negE.pmc         0.088                                                       
## psA_lg1.pmc     -0.020  0.024                                                
## dpgrp2rmttd     -0.859 -0.076   0.002                                        
## dpgrp2cntrl     -0.783 -0.070   0.005  0.673                                 
## negE.pm         -0.722  0.015  -0.002  0.621   0.566                         
## ngE.pmc:dpgrp2r -0.077 -0.869   0.000  0.083   0.061  -0.013                 
## ngE.pmc:dpgrp2c -0.067 -0.758  -0.001  0.058   0.077  -0.012   0.659         
## dpgrp2rm:E.      0.608 -0.013   0.000 -0.730  -0.476  -0.842   0.022         
## dpgrp2cn:E.      0.462 -0.010  -0.001 -0.397  -0.706  -0.639   0.009         
##                 ngE.pmc:dpgrp2c dpgrp2r:E.
## negE.pmc                                  
## psA_lg1.pmc                               
## dpgrp2rmttd                               
## dpgrp2cntrl                               
## negE.pm                                   
## ngE.pmc:dpgrp2r                           
## ngE.pmc:dpgrp2c                           
## dpgrp2rm:E.      0.010                    
## dpgrp2cn:E.      0.031           0.538

6.4.4 PA-reactivity to positive events

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: posA ~ posE.pmc + posA_lag1.pmc + depgroup2 + posE.pm + depgroup2 *  
##     posE.pmc + depgroup2 * posE.pm + (posE.pmc + posA_lag1.pmc |      pident)
##    Data: emaD
## 
## REML criterion at convergence: 33841.9
## 
## Scaled residuals: 
##    Min     1Q Median     3Q    Max 
## -6.064 -0.479  0.091  0.576  3.825 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr       
##  pident   (Intercept)   0.4849   0.696               
##           posE.pmc      0.0503   0.224    -0.44      
##           posA_lag1.pmc 0.0191   0.138    -0.04 -0.14
##  Residual               0.3902   0.625               
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##                            Estimate Std. Error       df t value Pr(>|t|)    
## (Intercept)                  2.7279     0.1479 358.0339   18.44  < 2e-16 ***
## posE.pmc                     0.4302     0.0415 327.9437   10.36  < 2e-16 ***
## posA_lag1.pmc                0.3149     0.0107 342.8789   29.33  < 2e-16 ***
## depgroup2remitted            0.8910     0.1734 357.8377    5.14  9.8e-07 ***
## depgroup2control             1.5700     0.1895 358.1907    8.28  6.8e-15 ***
## posE.pm                      1.0965     0.4149 351.3639    2.64  0.01268 *  
## posE.pmc:depgroup2remitted  -0.0634     0.0474 319.2013   -1.34  0.21074    
## posE.pmc:depgroup2control   -0.1886     0.0532 326.7822   -3.55  0.00077 ***
## depgroup2remitted:posE.pm   -0.7161     0.4682 350.8748   -1.53  0.14876    
## depgroup2control:posE.pm    -0.9994     0.4965 351.8726   -2.01  0.05903 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##                 (Intr) psE.pmc psA_1. dpgrp2r dpgrp2c posE.pm psE.pmc:dpgrp2r
## posE.pmc        -0.161                                                       
## psA_lg1.pmc     -0.007 -0.064                                                
## dpgrp2rmttd     -0.853  0.138   0.001                                        
## dpgrp2cntrl     -0.780  0.126   0.001  0.665                                 
## posE.pm         -0.794 -0.034  -0.002  0.677   0.620                         
## psE.pmc:dpgrp2r  0.142 -0.873   0.013 -0.157  -0.111   0.030                 
## psE.pmc:dpgrp2c  0.126 -0.779   0.016 -0.108  -0.158   0.027   0.682         
## dpgrp2rm:E.      0.704  0.030   0.000 -0.800  -0.549  -0.886  -0.037         
## dpgrp2cn:E.      0.664  0.028   0.001 -0.566  -0.791  -0.836  -0.025         
##                 psE.pmc:dpgrp2c dpgrp2r:E.
## posE.pmc                                  
## psA_lg1.pmc                               
## dpgrp2rmttd                               
## dpgrp2cntrl                               
## posE.pm                                   
## psE.pmc:dpgrp2r                           
## psE.pmc:dpgrp2c                           
## dpgrp2rm:E.     -0.024                    
## dpgrp2cn:E.     -0.037           0.740

7 Analysis step 4: Residual plots for replication models (DHARMa package)

7.1 NA-reactivity to negative events

DHARMa_simres <- simulateResiduals(fittedModel = Model_NA_negE_groups)
par(mfrow=c(1,2))
plotQQunif(DHARMa_simres, testUniformity = F, testOutliers = F, testDispersion = F)
plotResiduals(DHARMa_simres, rank = T, smoothScatter = F)

7.2 NA-reactivity to positive events

DHARMa_simres <- simulateResiduals(fittedModel = Model_NA_posE_groups)
par(mfrow=c(1,2))
plotQQunif(DHARMa_simres, testUniformity = F, testOutliers = F, testDispersion = F)
plotResiduals(DHARMa_simres, rank = T, smoothScatter = F)

7.3 PA-reactivity to negative events

DHARMa_simres <- simulateResiduals(fittedModel = Model_PA_negE_groups)
par(mfrow=c(1,2))
plotQQunif(DHARMa_simres, testUniformity = F, testOutliers = F, testDispersion = F)
plotResiduals(DHARMa_simres, rank = T, smoothScatter = F)

7.4 PA-reactivity to positive events

DHARMa_simres <- simulateResiduals(fittedModel = Model_PA_posE_groups)
par(mfrow=c(1,2))
plotQQunif(DHARMa_simres, testUniformity = F, testOutliers = F, testDispersion = F)
plotResiduals(DHARMa_simres, rank = T, smoothScatter = F)

8 Analysis step 5: Models of transformed negative affect and generalized linear multilevel models of negative affect

Analyses here are based on NA with its original scale (1-7), because the log transformation and GLMM outcome distributiuons only allow for positive values.

emaD$negA1_7 <- emaD$negA + 1
emaD$negA_log <- log(emaD$negA1_7)

8.1 Models of log-transformed negative affect

8.1.1 Negative events

Model_NAlog_negE <- lmer(negA_log ~ negE.pmc + negA_lag1.pmc + depgroup + negE.pm + depgroup*negE.pmc + depgroup*negE.pm + (negE.pmc + negA_lag1.pmc|pident), data = emaD)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negA_log ~ negE.pmc + negA_lag1.pmc + depgroup + negE.pm + depgroup *  
##     negE.pmc + depgroup * negE.pm + (negE.pmc + negA_lag1.pmc |      pident)
##    Data: emaD
## 
## REML criterion at convergence: -2489.8
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -4.9864 -0.5191 -0.1143  0.4331  6.5841 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr       
##  pident   (Intercept)   0.085072 0.2917              
##           negE.pmc      0.021715 0.1474   -0.01      
##           negA_lag1.pmc 0.005686 0.0754   -0.10  0.03
##  Residual               0.044265 0.2104              
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##                             Estimate Std. Error         df t value Pr(>|t|)    
## (Intercept)                 0.122360   0.043264 340.675255   2.828  0.00496 ** 
## negE.pmc                    0.168104   0.020653 286.376415   8.140 1.23e-14 ***
## negA_lag1.pmc               0.165348   0.005978 273.317741  27.660  < 2e-16 ***
## depgroupremitted            0.174968   0.054102 340.460030   3.234  0.00134 ** 
## depgroupcurrent             0.668498   0.069482 339.691729   9.621  < 2e-16 ***
## negE.pm                     0.508891   0.328447 340.854707   1.549  0.12222    
## negE.pmc:depgroupremitted   0.048415   0.024793 275.181947   1.953  0.05186 .  
## negE.pmc:depgroupcurrent    0.060809   0.031718 271.723212   1.917  0.05627 .  
## depgroupremitted:negE.pm    0.043014   0.372253 340.616316   0.116  0.90808    
## depgroupcurrent:negE.pm    -0.938961   0.426404 340.159862  -2.202  0.02833 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##                (Intr) ngE.pmc ngA_1. dpgrpr dpgrpc negE.pm ngE.pmc:dpgrpr
## negE.pmc       -0.002                                                    
## ngA_lg1.pmc    -0.012 -0.004                                             
## depgrprmttd    -0.799  0.002  -0.010                                     
## depgrpcrrnt    -0.622  0.002  -0.014  0.498                              
## negE.pm        -0.723 -0.002  -0.004  0.578  0.450                       
## ngE.pmc:dpgrpr  0.002 -0.833   0.002 -0.002 -0.001  0.002                
## ngE.pmc:dpgrpc  0.001 -0.651   0.006 -0.001 -0.002  0.001   0.542        
## dpgrprmt:E.     0.638  0.002   0.001 -0.723 -0.397 -0.882  -0.002        
## dpgrpcrr:E.     0.557  0.002   0.002 -0.445 -0.708 -0.770  -0.001        
##                ngE.pmc:dpgrpc dpgrpr:E.
## negE.pmc                               
## ngA_lg1.pmc                            
## depgrprmttd                            
## depgrpcrrnt                            
## negE.pm                                
## ngE.pmc:dpgrpr                         
## ngE.pmc:dpgrpc                         
## dpgrprmt:E.    -0.001                  
## dpgrpcrr:E.    -0.001          0.680

8.1.1.1 DHARMa residual plots

DHARMa_simres <- simulateResiduals(fittedModel = Model_NAlog_negE)
par(mfrow=c(1,2))
plotQQunif(DHARMa_simres, testUniformity = F, testOutliers = F, testDispersion = F)
plotResiduals(DHARMa_simres, rank = T, smoothScatter = F)

8.1.2 Positive events

Model_NAlog_posE <- lmer(negA_log ~ posE.pmc + negA_lag1.pmc + depgroup + posE.pm + depgroup*posE.pmc + depgroup*posE.pm + (posE.pmc + negA_lag1.pmc|pident), data = emaD)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: negA_log ~ posE.pmc + negA_lag1.pmc + depgroup + posE.pm + depgroup *  
##     posE.pmc + depgroup * posE.pm + (posE.pmc + negA_lag1.pmc |      pident)
##    Data: emaD
## 
## REML criterion at convergence: -1453.4
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -4.4280 -0.5270 -0.1314  0.4086  6.2140 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr       
##  pident   (Intercept)   0.087678 0.29610             
##           posE.pmc      0.004933 0.07024  -0.43      
##           negA_lag1.pmc 0.005045 0.07103  -0.13 -0.03
##  Residual               0.047697 0.21840             
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##                             Estimate Std. Error         df t value Pr(>|t|)    
## (Intercept)                 0.211788   0.050403 355.236272   4.202 3.35e-05 ***
## posE.pmc                   -0.044600   0.011095 308.193155  -4.020 7.33e-05 ***
## negA_lag1.pmc               0.166312   0.005861 278.058670  28.378  < 2e-16 ***
## depgroupremitted            0.218848   0.063412 355.033622   3.451 0.000625 ***
## depgroupcurrent             0.634751   0.080464 354.199714   7.889 3.84e-14 ***
## posE.pm                    -0.119140   0.116337 351.329269  -1.024 0.306497    
## posE.pmc:depgroupremitted  -0.045857   0.013474 299.037658  -3.403 0.000757 ***
## posE.pmc:depgroupcurrent   -0.060798   0.017748 309.104730  -3.426 0.000696 ***
## depgroupremitted:posE.pm   -0.042919   0.148609 350.368619  -0.289 0.772900    
## depgroupcurrent:posE.pm    -0.300872   0.211328 349.454420  -1.424 0.155420    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##                (Intr) psE.pmc ngA_1. dpgrpr dpgrpc posE.pm psE.pmc:dpgrpr
## posE.pmc       -0.138                                                    
## ngA_lg1.pmc    -0.019  0.006                                             
## depgrprmttd    -0.794  0.109  -0.010                                     
## depgrpcrrnt    -0.626  0.086  -0.012  0.498                              
## posE.pm        -0.798 -0.039   0.005  0.635  0.500                       
## psE.pmc:dpgrpr  0.113 -0.823  -0.002 -0.135 -0.071  0.032                
## psE.pmc:dpgrpc  0.086 -0.625   0.005 -0.068 -0.141  0.025   0.515        
## dpgrprmt:E.     0.625  0.031  -0.001 -0.809 -0.391 -0.783  -0.040        
## dpgrpcrr:E.     0.440  0.022  -0.004 -0.349 -0.793 -0.551  -0.018        
##                psE.pmc:dpgrpc dpgrpr:E.
## posE.pmc                               
## ngA_lg1.pmc                            
## depgrprmttd                            
## depgrpcrrnt                            
## posE.pm                                
## psE.pmc:dpgrpr                         
## psE.pmc:dpgrpc                         
## dpgrprmt:E.    -0.019                  
## dpgrpcrr:E.    -0.036          0.431

8.1.2.1 DHARMa residual plots

DHARMa_simres <- simulateResiduals(fittedModel = Model_NAlog_posE)
par(mfrow=c(1,2))
plotQQunif(DHARMa_simres, testUniformity = F, testOutliers = F, testDispersion = F)
plotResiduals(DHARMa_simres, rank = T, smoothScatter = F)

8.2 Generalized linear multilevel model - Gamma distribution

8.2.1 Negative events

GLMM_NA_negE_gamma <- glmer(negA1_7 ~ negE.pmc + negA_lag1.pmc + depgroup + negE.pm + depgroup*negE.pmc + depgroup*negE.pm + (negE.pmc + negA_lag1.pmc|pident),
                             data = emaD,
                             family = Gamma(link = "log"), control=glmerControl(optimizer = "bobyqa",optCtrl=list(maxfun=2e6)))

## Generalized linear mixed model fit by maximum likelihood (Laplace
##   Approximation) [glmerMod]
##  Family: Gamma  ( log )
## Formula: negA1_7 ~ negE.pmc + negA_lag1.pmc + depgroup + negE.pm + depgroup *  
##     negE.pmc + depgroup * negE.pm + (negE.pmc + negA_lag1.pmc |      pident)
##    Data: emaD
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 2e+06))
## 
##      AIC      BIC   logLik deviance df.resid 
##   9962.9  10094.4  -4964.5   9928.9    16828 
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.8931 -0.5255 -0.1626  0.3358 11.7424 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr       
##  pident   (Intercept)   0.01730  0.13152             
##           negE.pmc      0.02137  0.14619   0.00      
##           negA_lag1.pmc 0.00611  0.07817  -0.04  0.06
##  Residual               0.05683  0.23839             
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##                            Estimate Std. Error t value Pr(>|z|)    
## (Intercept)                0.126007   0.054185   2.326  0.02004 *  
## negE.pmc                   0.178979   0.029356   6.097 1.08e-09 ***
## negA_lag1.pmc              0.159458   0.008663  18.406  < 2e-16 ***
## depgroupremitted           0.188267   0.068284   2.757  0.00583 ** 
## depgroupcurrent            0.681720   0.088301   7.720 1.16e-14 ***
## negE.pm                    0.582243   0.407414   1.429  0.15297    
## negE.pmc:depgroupremitted  0.060630   0.035761   1.695  0.08999 .  
## negE.pmc:depgroupcurrent   0.046426   0.046076   1.008  0.31365    
## depgroupremitted:negE.pm  -0.031331   0.464702  -0.067  0.94625    
## depgroupcurrent:negE.pm   -0.935581   0.530541  -1.763  0.07783 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##                (Intr) ngE.pmc ngA_1. dpgrpr dpgrpc negE.pm ngE.pmc:dpgrpr
## negE.pmc        0.002                                                    
## ngA_lg1.pmc    -0.027  0.016                                             
## depgrprmttd    -0.797 -0.002   0.011                                     
## depgrpcrrnt    -0.619 -0.002   0.011  0.499                              
## negE.pm        -0.720 -0.006   0.008  0.573  0.445                       
## ngE.pmc:dpgrpr -0.001 -0.820   0.011  0.001  0.001  0.005                
## ngE.pmc:dpgrpc -0.002 -0.640   0.018  0.001 -0.002  0.005   0.525        
## dpgrprmt:E.     0.634  0.006  -0.015 -0.724 -0.398 -0.878  -0.006        
## dpgrpcrr:E.     0.555  0.005  -0.011 -0.447 -0.713 -0.768  -0.004        
##                ngE.pmc:dpgrpc dpgrpr:E.
## negE.pmc                               
## ngA_lg1.pmc                            
## depgrprmttd                            
## depgrpcrrnt                            
## negE.pm                                
## ngE.pmc:dpgrpr                         
## ngE.pmc:dpgrpc                         
## dpgrprmt:E.    -0.004                  
## dpgrpcrr:E.    -0.005          0.680

8.2.1.1 DHARMa residual plots

DHARMa_simres <- simulateResiduals(fittedModel = GLMM_NA_negE_gamma)
par(mfrow=c(1,2))
plotQQunif(DHARMa_simres, testUniformity = F, testOutliers = F, testDispersion = F)
plotResiduals(DHARMa_simres, rank = T, smoothScatter = F)

8.2.2 Positive events

GLMM_NA_posE_gamma <- glmer(negA1_7 ~ posE.pmc + negA_lag1.pmc + depgroup + posE.pm + depgroup*posE.pmc + depgroup*posE.pm + (posE.pmc + negA_lag1.pmc|pident),
                             data = emaD,
                             family = Gamma(link = "log"), control=glmerControl(optimizer = "bobyqa",optCtrl=list(maxfun=2e6)))

## Generalized linear mixed model fit by maximum likelihood (Laplace
##   Approximation) [glmerMod]
##  Family: Gamma  ( log )
## Formula: negA1_7 ~ posE.pmc + negA_lag1.pmc + depgroup + posE.pm + depgroup *  
##     posE.pmc + depgroup * posE.pm + (posE.pmc + negA_lag1.pmc |      pident)
##    Data: emaD
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 2e+06))
## 
##      AIC      BIC   logLik deviance df.resid 
##  11411.3  11542.7  -5688.6  11377.3    16828 
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.5256 -0.5148 -0.1764  0.3031 10.4184 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr       
##  pident   (Intercept)   0.019076 0.13811             
##           posE.pmc      0.007949 0.08916  -0.17      
##           negA_lag1.pmc 0.005832 0.07637  -0.05  0.01
##  Residual               0.064183 0.25334             
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##                            Estimate Std. Error t value Pr(>|z|)    
## (Intercept)                0.221574   0.062894   3.523 0.000427 ***
## posE.pmc                  -0.043409   0.016605  -2.614 0.008941 ** 
## negA_lag1.pmc              0.159042   0.008442  18.840  < 2e-16 ***
## depgroupremitted           0.222373   0.080089   2.777 0.005494 ** 
## depgroupcurrent            0.647278   0.100677   6.429 1.28e-10 ***
## posE.pm                   -0.116278   0.146061  -0.796 0.425978    
## posE.pmc:depgroupremitted -0.051657   0.020337  -2.540 0.011083 *  
## posE.pmc:depgroupcurrent  -0.063992   0.026655  -2.401 0.016360 *  
## depgroupremitted:posE.pm  -0.029251   0.188803  -0.155 0.876879    
## depgroupcurrent:posE.pm   -0.278801   0.265951  -1.048 0.294493    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##                (Intr) psE.pmc ngA_1. dpgrpr dpgrpc posE.pm psE.pmc:dpgrpr
## posE.pmc       -0.138                                                    
## ngA_lg1.pmc    -0.022  0.013                                             
## depgrprmttd    -0.786  0.107   0.011                                     
## depgrpcrrnt    -0.625  0.086   0.000  0.493                              
## posE.pm        -0.798 -0.018   0.004  0.626  0.499                       
## psE.pmc:dpgrpr  0.112 -0.817   0.000 -0.130 -0.070  0.014                
## psE.pmc:dpgrpc  0.086 -0.625  -0.007 -0.068 -0.144  0.011   0.510        
## dpgrprmt:E.     0.619  0.015  -0.017 -0.813 -0.388 -0.773  -0.021        
## dpgrpcrr:E.     0.438  0.010   0.000 -0.342 -0.794 -0.549  -0.008        
##                psE.pmc:dpgrpc dpgrpr:E.
## posE.pmc                               
## ngA_lg1.pmc                            
## depgrprmttd                            
## depgrpcrrnt                            
## posE.pm                                
## psE.pmc:dpgrpr                         
## psE.pmc:dpgrpc                         
## dpgrprmt:E.    -0.008                  
## dpgrpcrr:E.    -0.008          0.423

8.2.2.1 DHARMa residual plots

DHARMa_simres <- simulateResiduals(fittedModel = GLMM_NA_posE_gamma)
par(mfrow=c(1,2))
plotQQunif(DHARMa_simres, testUniformity = F, testOutliers = F, testDispersion = F)
plotResiduals(DHARMa_simres, rank = T, smoothScatter = F)

8.3 Generalized linear multilevel model - inverse Gaussian distribution

Note that models here deviated from other models by including log-transformed, non-centered negative affect at lag1. This was done to overcome convergence issues.

emaD$negA_lag1_log <-  log(emaD$negA_lag1) 

8.3.1 Negative events

GLMM_NA_negE_invG <- glmer(negA1_7 ~ negE.pmc + negA_lag1_log + depgroup + negE.pm + depgroup*negE.pmc + depgroup*negE.pm + (negE.pmc + negA_lag1_log|pident),
                             data = emaD,
                             family = inverse.gaussian(link = "log"), control=glmerControl(optimizer = "bobyqa",optCtrl=list(maxfun=2e6)))

## Generalized linear mixed model fit by maximum likelihood (Laplace
##   Approximation) [glmerMod]
##  Family: inverse.gaussian  ( log )
## Formula: negA1_7 ~ negE.pmc + negA_lag1.pmc + depgroup + negE.pm + depgroup *  
##     negE.pmc + depgroup * negE.pm + (negE.pmc + negA_lag1.pmc |      pident)
##    Data: emaD
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 2e+06))
## 
##      AIC      BIC   logLik deviance df.resid 
##   8130.2   8261.7  -4048.1   8096.2    16828 
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -2.0059 -0.5489 -0.1651  0.3497 14.0448 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr     
##  pident   (Intercept)   0.018043 0.13432           
##           negE.pmc      0.021415 0.14634  0.02     
##           negA_lag1.pmc 0.007607 0.08722  0.01 0.09
##  Residual               0.036619 0.19136           
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##                             Estimate Std. Error   t value Pr(>|z|)    
## (Intercept)                0.1736182  0.0002719   638.473   <2e-16 ***
## negE.pmc                   0.1778413  0.0002717   654.650   <2e-16 ***
## negA_lag1.pmc              0.1659072  0.0002718   610.394   <2e-16 ***
## depgroupremitted           0.1905331  0.0002720   700.376   <2e-16 ***
## depgroupcurrent            0.6632634  0.0002720  2438.301   <2e-16 ***
## negE.pm                    0.5000788  0.0002721  1837.946   <2e-16 ***
## negE.pmc:depgroupremitted  0.0605574  0.0002717   222.845   <2e-16 ***
## negE.pmc:depgroupcurrent   0.0659128  0.0353283     1.866   0.0621 .  
## depgroupremitted:negE.pm   0.1003831  0.0002721   368.907   <2e-16 ***
## depgroupcurrent:negE.pm   -0.8724674  0.0002721 -3206.472   <2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##                (Intr) ngE.pmc ngA_1. dpgrpr dpgrpc negE.pm ngE.pmc:dpgrpr
## negE.pmc       -0.001                                                    
## ngA_lg1.pmc     0.000  0.001                                             
## depgrprmttd     0.000  0.000   0.000                                     
## depgrpcrrnt     0.000  0.000   0.000  0.000                              
## negE.pm         0.000  0.000   0.000  0.000  0.000                       
## ngE.pmc:dpgrpr  0.000 -0.001   0.001  0.000  0.000  0.000                
## ngE.pmc:dpgrpc  0.013  0.005  -0.012  0.013  0.013  0.013   0.013        
## dpgrprmt:E.     0.000  0.000   0.000  0.000  0.000  0.000   0.000        
## dpgrpcrr:E.     0.000  0.000   0.000  0.000  0.000  0.000   0.000        
##                ngE.pmc:dpgrpc dpgrpr:E.
## negE.pmc                               
## ngA_lg1.pmc                            
## depgrprmttd                            
## depgrpcrrnt                            
## negE.pm                                
## ngE.pmc:dpgrpr                         
## ngE.pmc:dpgrpc                         
## dpgrprmt:E.     0.013                  
## dpgrpcrr:E.     0.013          0.000   
## optimizer (bobyqa) convergence code: 0 (OK)
## Model failed to converge with max|grad| = 0.984993 (tol = 0.002, component 1)
## Model is nearly unidentifiable: very large eigenvalue
##  - Rescale variables?

8.3.1.1 DHARMa residual plots

DHARMa_simres <- simulateResiduals(fittedModel = GLMM_NA_negE_invG)
par(mfrow=c(1,2))
plotQQunif(DHARMa_simres, testUniformity = F, testOutliers = F, testDispersion = F)
plotResiduals(DHARMa_simres, rank = T, smoothScatter = F)

8.3.2 Positive events

GLMM_NA_posE_invG <- glmer(negA1_7 ~ posE.pmc + negA_lag1_log + depgroup + posE.pm + depgroup*posE.pmc + depgroup*posE.pm + (posE.pmc + negA_lag1_log|pident),
                             data = emaD,
                             family = inverse.gaussian(link = "log"), control=glmerControl(optimizer = "bobyqa",optCtrl=list(maxfun=2e6)))

## Generalized linear mixed model fit by maximum likelihood (Laplace
##   Approximation) [glmerMod]
##  Family: inverse.gaussian  ( log )
## Formula: negA1_7 ~ posE.pmc + negA_lag1.pmc + depgroup + posE.pm + depgroup *  
##     posE.pmc + depgroup * posE.pm + (posE.pmc + negA_lag1.pmc |      pident)
##    Data: emaD
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 2e+06))
## 
##      AIC      BIC   logLik deviance df.resid 
##   9570.6   9702.0  -4768.3   9536.6    16828 
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -1.8737 -0.5344 -0.1811  0.3012 10.7934 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr       
##  pident   (Intercept)   0.019798 0.14071             
##           posE.pmc      0.007709 0.08780  -0.20      
##           negA_lag1.pmc 0.007508 0.08665   0.00  0.00
##  Residual               0.041737 0.20430             
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##                             Estimate Std. Error  t value Pr(>|z|)    
## (Intercept)                0.2478722  0.0003700  669.886  < 2e-16 ***
## posE.pmc                  -0.0469763  0.0003700 -126.949  < 2e-16 ***
## negA_lag1.pmc              0.1576869  0.0003700  426.197  < 2e-16 ***
## depgroupremitted           0.2291623  0.0003823  599.382  < 2e-16 ***
## depgroupcurrent            0.6689494  0.0003702 1806.916  < 2e-16 ***
## posE.pm                   -0.1209126  0.0003703 -326.555  < 2e-16 ***
## posE.pmc:depgroupremitted -0.0555084  0.0110734   -5.013 5.37e-07 ***
## posE.pmc:depgroupcurrent  -0.0706459  0.0003702 -190.851  < 2e-16 ***
## depgroupremitted:posE.pm   0.0469178  0.0003824  122.681  < 2e-16 ***
## depgroupcurrent:posE.pm   -0.2675140  0.0003825 -699.456  < 2e-16 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##                (Intr) psE.pmc ngA_1. dpgrpr dpgrpc posE.pm psE.pmc:dpgrpr
## posE.pmc        0.000                                                    
## ngA_lg1.pmc     0.001  0.000                                             
## depgrprmttd     0.000  0.000   0.000                                     
## depgrpcrrnt     0.000  0.000   0.000  0.000                              
## posE.pm         0.000  0.000   0.000  0.000  0.000                       
## psE.pmc:dpgrpr -0.008 -0.037   0.004 -0.007 -0.004 -0.005                
## psE.pmc:dpgrpc  0.000  0.000   0.000  0.000  0.000  0.000  -0.004        
## dpgrprmt:E.     0.000  0.000   0.000  0.000  0.000  0.000  -0.004        
## dpgrpcrr:E.     0.000  0.000   0.000  0.000  0.000  0.000  -0.002        
##                psE.pmc:dpgrpc dpgrpr:E.
## posE.pmc                               
## ngA_lg1.pmc                            
## depgrprmttd                            
## depgrpcrrnt                            
## posE.pm                                
## psE.pmc:dpgrpr                         
## psE.pmc:dpgrpc                         
## dpgrprmt:E.     0.000                  
## dpgrpcrr:E.     0.000         -0.250   
## optimizer (bobyqa) convergence code: 0 (OK)
## Model failed to converge with max|grad| = 0.197621 (tol = 0.002, component 1)
## Model is nearly unidentifiable: very large eigenvalue
##  - Rescale variables?

8.3.2.1 DHARMa residual plots

DHARMa_simres <- simulateResiduals(fittedModel = GLMM_NA_posE_invG)
par(mfrow=c(1,2))
plotQQunif(DHARMa_simres, testUniformity = F, testOutliers = F, testDispersion = F)
plotResiduals(DHARMa_simres, rank = T, smoothScatter = F)

9 Analysis step 5: Two-part models of negative affect

9.1 Random intercept only

9.1.1 Negative events

NA_negE_twopart_1 <- mixed_model(fixed=negA ~  negE.pmc + negA_lag1.pmc + depgroup + negE.pm + depgroup*negE.pmc + depgroup*negE.pm, random=~1 |pident,
                               data = emaD,
                               zi_fixed = ~ negE.pmc + negA_lag1.pmc + depgroup + negE.pm + depgroup*negE.pmc + depgroup*negE.pm,
                               zi_random= NULL,
                               family = hurdle.lognormal(), n_phis=1, control=(iter_EM = 0))

9.1.1.1 model overview

## $logLik
## 'log Lik.' -20036.58 (df=22)
## 
## $AIC
## [1] 40117.15
## 
## $BIC
## [1] 40201.77
## 
## $call
## mixed_model(fixed = negA ~ negE.pmc + negA_lag1.pmc + depgroup + 
##     negE.pm + depgroup * negE.pmc + depgroup * negE.pm, random = ~1 | 
##     pident, data = emaD, family = hurdle.lognormal(), zi_fixed = ~negE.pmc + 
##     negA_lag1.pmc + depgroup + negE.pm + depgroup * negE.pmc + 
##     depgroup * negE.pm, zi_random = NULL, n_phis = 1, control = (iter_EM = 0))
## 
## $N
## [1] 16845
## 
## $phis_table
##        Estimate     Std.Err
## phi_1 -0.508282 0.006967357
## 
## $family
## 
## Family: two-part log-normal 
## Link function: identity

9.1.1.2 coefficients for the continuous part

##                           Estimate Std.Err  z-value p-value
## (Intercept)                -1.2198  0.0978 -12.4754  0.0000
## negE.pmc                    0.3986  0.0409   9.7487  0.0000
## negA_lag1.pmc               0.3400  0.0102  33.1858  0.0000
## depgroupremitted            0.3810  0.1201   3.1739  0.0015
## depgroupcurrent             1.2949  0.1509   8.5837  0.0000
## negE.pm                     0.5796  0.7142   0.8116  0.4170
## negE.pmc:depgroupremitted  -0.0007  0.0465  -0.0146  0.9883
## negE.pmc:depgroupcurrent    0.0260  0.0556   0.4678  0.6400
## depgroupremitted:negE.pm    0.3172  0.8057   0.3936  0.6939
## depgroupcurrent:negE.pm    -1.3544  0.9168  -1.4772  0.1396

9.1.1.3 coefficients for the zero-inflated part

##                           Estimate Std.Err  z-value p-value
## (Intercept)                 0.6570  0.0459  14.3183  0.0000
## negE.pmc                   -1.0438  0.1145  -9.1150  0.0000
## negA_lag1.pmc              -0.6076  0.0415 -14.6492  0.0000
## depgroupremitted           -0.9114  0.0574 -15.8772  0.0000
## depgroupcurrent            -2.9336  0.0930 -31.5505  0.0000
## negE.pm                    -3.8045  0.3683 -10.3305  0.0000
## negE.pmc:depgroupremitted   0.1222  0.1432   0.8535  0.3934
## negE.pmc:depgroupcurrent    0.1912  0.2598   0.7358  0.4619
## depgroupremitted:negE.pm    0.5488  0.4267   1.2863  0.1983
## depgroupcurrent:negE.pm     5.0871  0.5087   9.9996  0.0000

9.1.1.4 exponential of coefficients for the zero-inflated part

##               (Intercept)                  negE.pmc             negA_lag1.pmc 
##                    1.9290                    0.3521                    0.5447 
##          depgroupremitted           depgroupcurrent                   negE.pm 
##                    0.4019                    0.0532                    0.0223 
## negE.pmc:depgroupremitted  negE.pmc:depgroupcurrent  depgroupremitted:negE.pm 
##                    1.1300                    1.2107                    1.7312 
##   depgroupcurrent:negE.pm 
##                  161.9253

9.1.1.5 random effects (intercept and slope (co)variance)

## $D
##             (Intercept)
## (Intercept)      0.3725
## attr(,"L")
## [1] 0.6103

9.1.2 Positive events

NA_posE_twopart_1 <- mixed_model(fixed=negA ~  posE.pmc + negA_lag1.pmc + depgroup + posE.pm + depgroup*posE.pmc + depgroup*posE.pm, random=~1 |pident,
                               data = emaD,
                               zi_fixed = ~ posE.pmc + negA_lag1.pmc + depgroup + posE.pm + depgroup*posE.pmc + depgroup*posE.pm,
                               zi_random= NULL,
                               family = hurdle.lognormal(), n_phis=1, control=(iter_EM = 0))

9.1.2.1 model overview

## $logLik
## 'log Lik.' -20431.54 (df=22)
## 
## $AIC
## [1] 40907.07
## 
## $BIC
## [1] 40991.69
## 
## $call
## mixed_model(fixed = negA ~ posE.pmc + negA_lag1.pmc + depgroup + 
##     posE.pm + depgroup * posE.pmc + depgroup * posE.pm, random = ~1 | 
##     pident, data = emaD, family = hurdle.lognormal(), zi_fixed = ~posE.pmc + 
##     negA_lag1.pmc + depgroup + posE.pm + depgroup * posE.pmc + 
##     depgroup * posE.pm, zi_random = NULL, n_phis = 1, control = (iter_EM = 0))
## 
## $N
## [1] 16845
## 
## $phis_table
##         Estimate     Std.Err
## phi_1 -0.4902802 0.006966042
## 
## $family
## 
## Family: two-part log-normal 
## Link function: identity

9.1.2.2 coefficients for the continuous part

##                           Estimate Std.Err z-value p-value
## (Intercept)                -0.9548  0.1128 -8.4686  0.0000
## posE.pmc                   -0.1090  0.0360 -3.0281  0.0025
## negA_lag1.pmc               0.3372  0.0105 32.2579  0.0000
## depgroupremitted            0.3203  0.1395  2.2963  0.0217
## depgroupcurrent             1.1270  0.1733  6.5021  0.0000
## posE.pm                    -0.5049  0.2649 -1.9061  0.0566
## posE.pmc:depgroupremitted  -0.1012  0.0407 -2.4877  0.0129
## posE.pmc:depgroupcurrent   -0.0803  0.0471 -1.7075  0.0877
## depgroupremitted:posE.pm    0.3230  0.3315  0.9743  0.3299
## depgroupcurrent:posE.pm    -0.2283  0.4589 -0.4976  0.6188

9.1.2.3 coefficients for the zero-inflated part

##                           Estimate Std.Err  z-value p-value
## (Intercept)                 0.1458  0.0509   2.8623  0.0042
## posE.pmc                    0.5496  0.0758   7.2488  0.0000
## negA_lag1.pmc              -0.5775  0.0405 -14.2651  0.0000
## depgroupremitted           -0.8204  0.0652 -12.5917  0.0000
## depgroupcurrent            -2.6642  0.1134 -23.5007  0.0000
## posE.pm                     0.4348  0.1197   3.6314  0.0003
## posE.pmc:depgroupremitted  -0.1522  0.0920  -1.6537  0.0982
## posE.pmc:depgroupcurrent   -0.0991  0.1535  -0.6455  0.5186
## depgroupremitted:posE.pm   -0.4137  0.1552  -2.6658  0.0077
## depgroupcurrent:posE.pm     1.0075  0.2753   3.6595  0.0003

9.1.2.4 exponential of coefficients for the zero-inflated part

##               (Intercept)                  posE.pmc             negA_lag1.pmc 
##                    1.1570                    1.7325                    0.5613 
##          depgroupremitted           depgroupcurrent                   posE.pm 
##                    0.4402                    0.0697                    1.5447 
## posE.pmc:depgroupremitted  posE.pmc:depgroupcurrent  depgroupremitted:posE.pm 
##                    0.8588                    0.9057                    0.6612 
##   depgroupcurrent:posE.pm 
##                    2.7387

9.1.2.5 random effects (intercept SD and variance)

## $D
##             (Intercept)
## (Intercept)      0.3605
## attr(,"L")
## [1] 0.6004

9.2 Random intercept and event slope

9.2.1 Negative events

NA_negE_twopart_2 <- mixed_model(fixed=negA ~  negE.pmc + negA_lag1.pmc + depgroup + negE.pm + depgroup*negE.pmc + depgroup*negE.pm, random=~1 + negE.pmc|pident,
                               data = emaD,
                               zi_fixed = ~ negE.pmc + negA_lag1.pmc + depgroup + negE.pm + depgroup*negE.pmc + depgroup*negE.pm,
                               zi_random= NULL,
                               family = hurdle.lognormal(), n_phis=1, control=(iter_EM = 0))

9.2.1.1 model overview

## $logLik
## 'log Lik.' -20005.3 (df=24)
## 
## $AIC
## [1] 40058.59
## 
## $BIC
## [1] 40150.91
## 
## $call
## mixed_model(fixed = negA ~ negE.pmc + negA_lag1.pmc + depgroup + 
##     negE.pm + depgroup * negE.pmc + depgroup * negE.pm, random = ~1 + 
##     negE.pmc | pident, data = emaD, family = hurdle.lognormal(), 
##     zi_fixed = ~negE.pmc + negA_lag1.pmc + depgroup + negE.pm + 
##         depgroup * negE.pmc + depgroup * negE.pm, zi_random = NULL, 
##     n_phis = 1, control = (iter_EM = 0))
## 
## $N
## [1] 16845
## 
## $phis_table
##         Estimate     Std.Err
## phi_1 -0.5187107 0.007068262
## 
## $family
## 
## Family: two-part log-normal 
## Link function: identity

9.2.1.2 coefficients for the continuous part

##                           Estimate Std.Err  z-value p-value
## (Intercept)                -1.2148  0.0972 -12.4947  0.0000
## negE.pmc                    0.4005  0.0531   7.5470  0.0000
## negA_lag1.pmc               0.3409  0.0102  33.3941  0.0000
## depgroupremitted            0.3959  0.1195   3.3123  0.0009
## depgroupcurrent             1.2978  0.1503   8.6319  0.0000
## negE.pm                     0.5582  0.7057   0.7910  0.4289
## negE.pmc:depgroupremitted   0.0316  0.0616   0.5126  0.6083
## negE.pmc:depgroupcurrent    0.0314  0.0745   0.4213  0.6735
## depgroupremitted:negE.pm    0.1651  0.7970   0.2072  0.8359
## depgroupcurrent:negE.pm    -1.3842  0.9080  -1.5244  0.1274

9.2.1.3 coefficients for the zero-inflated part

##                           Estimate Std.Err  z-value p-value
## (Intercept)                 0.6570  0.0459  14.3183  0.0000
## negE.pmc                   -1.0438  0.1145  -9.1150  0.0000
## negA_lag1.pmc              -0.6076  0.0415 -14.6492  0.0000
## depgroupremitted           -0.9114  0.0574 -15.8772  0.0000
## depgroupcurrent            -2.9336  0.0930 -31.5505  0.0000
## negE.pm                    -3.8045  0.3683 -10.3305  0.0000
## negE.pmc:depgroupremitted   0.1222  0.1432   0.8535  0.3934
## negE.pmc:depgroupcurrent    0.1912  0.2598   0.7358  0.4619
## depgroupremitted:negE.pm    0.5488  0.4267   1.2863  0.1983
## depgroupcurrent:negE.pm     5.0871  0.5087   9.9996  0.0000

9.2.1.4 exponential of coefficients for the zero-inflated part

##               (Intercept)                  negE.pmc             negA_lag1.pmc 
##                    1.9290                    0.3521                    0.5447 
##          depgroupremitted           depgroupcurrent                   negE.pm 
##                    0.4019                    0.0532                    0.0223 
## negE.pmc:depgroupremitted  negE.pmc:depgroupcurrent  depgroupremitted:negE.pm 
##                    1.1300                    1.2107                    1.7312 
##   depgroupcurrent:negE.pm 
##                  161.9253

9.2.1.5 random effects (intercept and slope (co)variance)

## $D
##             (Intercept) negE.pmc
## (Intercept)     0.37429 -0.04074
## negE.pmc       -0.04074  0.06733
## attr(,"L")
## [1]  0.61180 -0.06659  0.25079

9.2.1.6 likelihood ratio test

anova(NA_negE_twopart_1, NA_negE_twopart_2)

## 
##                        AIC      BIC   log.Lik   LRT df p.value
## NA_negE_twopart_1 40117.15 40201.77 -20036.58                 
## NA_negE_twopart_2 40058.59 40150.91 -20005.30 62.56  2 <0.0001

9.2.2 Positive events

NA_posE_twopart_2 <- mixed_model(fixed=negA ~  posE.pmc + negA_lag1.pmc + depgroup + posE.pm + depgroup*posE.pmc + depgroup*posE.pm, random=~1 + posE.pmc |pident,
                               data = emaD,
                               zi_fixed = ~ posE.pmc + negA_lag1.pmc + depgroup + posE.pm + depgroup*posE.pmc + depgroup*posE.pm,
                               zi_random= NULL,
                               family = hurdle.lognormal(), n_phis=1, control=(iter_EM = 0))

9.2.2.1 model overview

## $logLik
## 'log Lik.' -20415.88 (df=24)
## 
## $AIC
## [1] 40879.76
## 
## $BIC
## [1] 40972.07
## 
## $call
## mixed_model(fixed = negA ~ posE.pmc + negA_lag1.pmc + depgroup + 
##     posE.pm + depgroup * posE.pmc + depgroup * posE.pm, random = ~1 + 
##     posE.pmc | pident, data = emaD, family = hurdle.lognormal(), 
##     zi_fixed = ~posE.pmc + negA_lag1.pmc + depgroup + posE.pm + 
##         depgroup * posE.pmc + depgroup * posE.pm, zi_random = NULL, 
##     n_phis = 1, control = (iter_EM = 0))
## 
## $N
## [1] 16845
## 
## $phis_table
##         Estimate     Std.Err
## phi_1 -0.4976181 0.007062241
## 
## $family
## 
## Family: two-part log-normal 
## Link function: identity

9.2.2.2 coefficients for the continuous part

##                           Estimate Std.Err z-value p-value
## (Intercept)                -0.9692  0.1127 -8.5981  0.0000
## posE.pmc                   -0.0972  0.0436 -2.2308  0.0257
## negA_lag1.pmc               0.3363  0.0104 32.2404  0.0000
## depgroupremitted            0.3584  0.1406  2.5495  0.0108
## depgroupcurrent             1.1379  0.1730  6.5769  0.0000
## posE.pm                    -0.4675  0.2648 -1.7655  0.0775
## posE.pmc:depgroupremitted  -0.0932  0.0498 -1.8692  0.0616
## posE.pmc:depgroupcurrent   -0.0878  0.0587 -1.4961  0.1346
## depgroupremitted:posE.pm    0.2151  0.3355  0.6411  0.5215
## depgroupcurrent:posE.pm    -0.2562  0.4576 -0.5599  0.5755

9.2.2.3 coefficients for the zero-inflated part

##                           Estimate Std.Err  z-value p-value
## (Intercept)                 0.1458  0.0509   2.8623  0.0042
## posE.pmc                    0.5496  0.0758   7.2488  0.0000
## negA_lag1.pmc              -0.5775  0.0405 -14.2651  0.0000
## depgroupremitted           -0.8204  0.0652 -12.5917  0.0000
## depgroupcurrent            -2.6642  0.1134 -23.5007  0.0000
## posE.pm                     0.4348  0.1197   3.6314  0.0003
## posE.pmc:depgroupremitted  -0.1522  0.0920  -1.6537  0.0982
## posE.pmc:depgroupcurrent   -0.0991  0.1535  -0.6455  0.5186
## depgroupremitted:posE.pm   -0.4137  0.1552  -2.6658  0.0077
## depgroupcurrent:posE.pm     1.0075  0.2753   3.6595  0.0003

9.2.2.4 exponential of coefficients for the zero-inflated part

##               (Intercept)                  posE.pmc             negA_lag1.pmc 
##                    1.1570                    1.7325                    0.5613 
##          depgroupremitted           depgroupcurrent                   posE.pm 
##                    0.4402                    0.0697                    1.5447 
## posE.pmc:depgroupremitted  posE.pmc:depgroupcurrent  depgroupremitted:posE.pm 
##                    0.8588                    0.9057                    0.6612 
##   depgroupcurrent:posE.pm 
##                    2.7387

9.2.2.5 random effects (intercept SD and variance)

## $D
##             (Intercept) posE.pmc
## (Intercept)     0.36042 -0.01836
## posE.pmc       -0.01836  0.03385
## attr(,"L")
## [1]  0.60035 -0.03058  0.18142

9.2.2.6 likelihood ratio test

anova(NA_posE_twopart_1, NA_posE_twopart_2)

## 
##                        AIC      BIC   log.Lik   LRT df p.value
## NA_posE_twopart_1 40907.07 40991.69 -20431.54                 
## NA_posE_twopart_2 40879.76 40972.07 -20415.88 31.31  2 <0.0001

9.3 Random intercept, event slope, and negA_lag1 slope

9.3.1 Negative events

NA_negE_twopart_3 <- mixed_model(fixed=negA ~  negE.pmc + negA_lag1.pmc + depgroup + negE.pm + depgroup*negE.pmc + depgroup*negE.pm, random=~1 + negE.pmc + negA_lag1.pmc|pident,
                               data = emaD,
                               zi_fixed = ~ negE.pmc + negA_lag1.pmc + depgroup + negE.pm + depgroup*negE.pmc + depgroup*negE.pm,
                               zi_random= NULL,
                               family = hurdle.lognormal(), n_phis=1, control=(iter_EM = 0))

9.3.1.1 model overview

## $logLik
## 'log Lik.' -19948.63 (df=27)
## 
## $AIC
## [1] 39951.26
## 
## $BIC
## [1] 40055.11
## 
## $call
## mixed_model(fixed = negA ~ negE.pmc + negA_lag1.pmc + depgroup + 
##     negE.pm + depgroup * negE.pmc + depgroup * negE.pm, random = ~1 + 
##     negE.pmc + negA_lag1.pmc | pident, data = emaD, family = hurdle.lognormal(), 
##     zi_fixed = ~negE.pmc + negA_lag1.pmc + depgroup + negE.pm + 
##         depgroup * negE.pmc + depgroup * negE.pm, zi_random = NULL, 
##     n_phis = 1, control = (iter_EM = 0))
## 
## $N
## [1] 16845
## 
## $phis_table
##         Estimate     Std.Err
## phi_1 -0.5303796 0.007152237
## 
## $family
## 
## Family: two-part log-normal 
## Link function: identity

9.3.1.2 coefficients for the continuous part

##                           Estimate Std.Err  z-value p-value
## (Intercept)                -1.2359  0.0956 -12.9300  0.0000
## negE.pmc                    0.4040  0.0520   7.7734  0.0000
## negA_lag1.pmc               0.3711  0.0171  21.7025  0.0000
## depgroupremitted            0.4499  0.1173   3.8361  0.0001
## depgroupcurrent             1.3375  0.1445   9.2585  0.0000
## negE.pm                     0.8365  0.6892   1.2138  0.2248
## negE.pmc:depgroupremitted   0.0341  0.0603   0.5645  0.5724
## negE.pmc:depgroupcurrent    0.0390  0.0731   0.5334  0.5938
## depgroupremitted:negE.pm   -0.3400  0.7813  -0.4352  0.6634
## depgroupcurrent:negE.pm    -1.9486  0.8757  -2.2251  0.0261

9.3.1.3 coefficients for the zero-inflated part

##                           Estimate Std.Err  z-value p-value
## (Intercept)                 0.6570  0.0459  14.3183  0.0000
## negE.pmc                   -1.0438  0.1145  -9.1150  0.0000
## negA_lag1.pmc              -0.6076  0.0415 -14.6492  0.0000
## depgroupremitted           -0.9114  0.0574 -15.8772  0.0000
## depgroupcurrent            -2.9336  0.0930 -31.5505  0.0000
## negE.pm                    -3.8045  0.3683 -10.3305  0.0000
## negE.pmc:depgroupremitted   0.1222  0.1432   0.8535  0.3934
## negE.pmc:depgroupcurrent    0.1912  0.2598   0.7358  0.4619
## depgroupremitted:negE.pm    0.5488  0.4267   1.2863  0.1983
## depgroupcurrent:negE.pm     5.0871  0.5087   9.9996  0.0000

9.3.1.4 exponential of coefficients for the zero-inflated part

##               (Intercept)                  negE.pmc             negA_lag1.pmc 
##                    1.9290                    0.3521                    0.5447 
##          depgroupremitted           depgroupcurrent                   negE.pm 
##                    0.4019                    0.0532                    0.0223 
## negE.pmc:depgroupremitted  negE.pmc:depgroupcurrent  depgroupremitted:negE.pm 
##                    1.1300                    1.2107                    1.7312 
##   depgroupcurrent:negE.pm 
##                  161.9253

9.3.1.5 random effects (intercept and slope (co)variance)

## $D
##               (Intercept)  negE.pmc negA_lag1.pmc
## (Intercept)       0.38003 -0.040974     -0.047577
## negE.pmc         -0.04097  0.061364      0.003973
## negA_lag1.pmc    -0.04758  0.003973      0.026341
## attr(,"L")
## [1]  0.616467 -0.066466 -0.077177  0.238635 -0.004845  0.142693

9.3.1.6 likelihood ratio test

anova(NA_negE_twopart_2, NA_negE_twopart_3)

## 
##                        AIC      BIC   log.Lik    LRT df p.value
## NA_negE_twopart_2 40058.59 40150.91 -20005.30                  
## NA_negE_twopart_3 39951.26 40055.11 -19948.63 113.33  3 <0.0001

9.3.2 Positive events

NA_posE_twopart_3 <- mixed_model(fixed=negA ~  posE.pmc + negA_lag1.pmc + depgroup + posE.pm + depgroup*posE.pmc + depgroup*posE.pm, random=~1 + posE.pmc + negA_lag1.pmc |pident,
                               data = emaD,
                               zi_fixed = ~ posE.pmc + negA_lag1.pmc + depgroup + posE.pm + depgroup*posE.pmc + depgroup*posE.pm,
                               zi_random= NULL,
                               family = hurdle.lognormal(), n_phis=1, control=(iter_EM = 0))

9.3.2.1 model overview

## $logLik
## 'log Lik.' -20364.93 (df=27)
## 
## $AIC
## [1] 40783.86
## 
## $BIC
## [1] 40887.72
## 
## $call
## mixed_model(fixed = negA ~ posE.pmc + negA_lag1.pmc + depgroup + 
##     posE.pm + depgroup * posE.pmc + depgroup * posE.pm, random = ~1 + 
##     posE.pmc + negA_lag1.pmc | pident, data = emaD, family = hurdle.lognormal(), 
##     zi_fixed = ~posE.pmc + negA_lag1.pmc + depgroup + posE.pm + 
##         depgroup * posE.pmc + depgroup * posE.pm, zi_random = NULL, 
##     n_phis = 1, control = (iter_EM = 0))
## 
## $N
## [1] 16845
## 
## $phis_table
##         Estimate     Std.Err
## phi_1 -0.5082036 0.007144654
## 
## $family
## 
## Family: two-part log-normal 
## Link function: identity

9.3.2.2 coefficients for the continuous part

##                           Estimate Std.Err z-value p-value
## (Intercept)                -0.9842  0.1109 -8.8723  0.0000
## posE.pmc                   -0.0891  0.0429 -2.0774  0.0378
## negA_lag1.pmc               0.3573  0.0167 21.3798  0.0000
## depgroupremitted            0.3990  0.1375  2.9029  0.0037
## depgroupcurrent             1.1188  0.1672  6.6903  0.0000
## posE.pm                    -0.4007  0.2616 -1.5316  0.1256
## posE.pmc:depgroupremitted  -0.1022  0.0490 -2.0835  0.0372
## posE.pmc:depgroupcurrent   -0.0938  0.0578 -1.6223  0.1047
## depgroupremitted:posE.pm    0.0754  0.3291  0.2290  0.8189
## depgroupcurrent:posE.pm    -0.2527  0.4393 -0.5752  0.5651

9.3.2.3 coefficients for the zero-inflated part

##                           Estimate Std.Err  z-value p-value
## (Intercept)                 0.1458  0.0509   2.8623  0.0042
## posE.pmc                    0.5496  0.0758   7.2488  0.0000
## negA_lag1.pmc              -0.5775  0.0405 -14.2651  0.0000
## depgroupremitted           -0.8204  0.0652 -12.5917  0.0000
## depgroupcurrent            -2.6642  0.1134 -23.5007  0.0000
## posE.pm                     0.4348  0.1197   3.6314  0.0003
## posE.pmc:depgroupremitted  -0.1522  0.0920  -1.6537  0.0982
## posE.pmc:depgroupcurrent   -0.0991  0.1535  -0.6455  0.5186
## depgroupremitted:posE.pm   -0.4137  0.1552  -2.6658  0.0077
## depgroupcurrent:posE.pm     1.0075  0.2753   3.6595  0.0003

9.3.2.4 exponential of coefficients for the zero-inflated part

##               (Intercept)                  posE.pmc             negA_lag1.pmc 
##                    1.1570                    1.7325                    0.5613 
##          depgroupremitted           depgroupcurrent                   posE.pm 
##                    0.4402                    0.0697                    1.5447 
## posE.pmc:depgroupremitted  posE.pmc:depgroupcurrent  depgroupremitted:posE.pm 
##                    0.8588                    0.9057                    0.6612 
##   depgroupcurrent:posE.pm 
##                    2.7387

9.3.2.5 random effects (intercept SD and variance)

## $D
##               (Intercept)  posE.pmc negA_lag1.pmc
## (Intercept)       0.36338 -0.022996     -0.040259
## posE.pmc         -0.02300  0.031406     -0.004679
## negA_lag1.pmc    -0.04026 -0.004679      0.023684
## attr(,"L")
## [1]  0.60281 -0.03815 -0.06679  0.17306 -0.04176  0.13221

9.3.2.6 likelihood ratio test

anova(NA_posE_twopart_2, NA_posE_twopart_3)

## 
##                        AIC      BIC   log.Lik    LRT df p.value
## NA_posE_twopart_2 40879.76 40972.07 -20415.88                  
## NA_posE_twopart_3 40783.86 40887.72 -20364.93 101.89  3 <0.0001

9.4 Adding random intercept in zero-inflated model (final two-part models)

9.4.1 Negative events

NA_negE_twopart_4 <- mixed_model(fixed=negA ~  negE.pmc + negA_lag1.pmc + depgroup + negE.pm + depgroup*negE.pmc + depgroup*negE.pm, random=~1 + negE.pmc + negA_lag1.pmc|pident,
                               data = emaD,
                               zi_fixed = ~ negE.pmc + negA_lag1.pmc + depgroup + negE.pm + depgroup*negE.pmc + depgroup*negE.pm,
                               zi_random= ~ 1|pident,
                               family = hurdle.lognormal(), n_phis=1, control=(iter_EM = 0))

9.4.1.1 model overview

## $logLik
## 'log Lik.' -16566.49 (df=31)
## 
## $AIC
## [1] 33194.98
## 
## $BIC
## [1] 33314.22
## 
## $call
## mixed_model(fixed = negA ~ negE.pmc + negA_lag1.pmc + depgroup + 
##     negE.pm + depgroup * negE.pmc + depgroup * negE.pm, random = ~1 + 
##     negE.pmc + negA_lag1.pmc | pident, data = emaD, family = hurdle.lognormal(), 
##     zi_fixed = ~negE.pmc + negA_lag1.pmc + depgroup + negE.pm + 
##         depgroup * negE.pmc + depgroup * negE.pm, zi_random = ~1 | 
##         pident, n_phis = 1, control = (iter_EM = 0))
## 
## $N
## [1] 16845
## 
## $phis_table
##         Estimate     Std.Err
## phi_1 -0.5303593 0.007147102
## 
## $family
## 
## Family: two-part log-normal 
## Link function: identity

9.4.1.2 coefficients for the continuous part

##                           Estimate Std.Err  z-value p-value
## (Intercept)                -1.3138  0.0937 -14.0250  0.0000
## negE.pmc                    0.4260  0.0522   8.1582  0.0000
## negA_lag1.pmc               0.3742  0.0175  21.3855  0.0000
## depgroupremitted            0.4877  0.1155   4.2231  0.0000
## depgroupcurrent             1.3584  0.1432   9.4867  0.0000
## negE.pm                     0.8280  0.6840   1.2106  0.2261
## negE.pmc:depgroupremitted   0.0223  0.0601   0.3707  0.7109
## negE.pmc:depgroupcurrent    0.0184  0.0731   0.2519  0.8011
## depgroupremitted:negE.pm   -0.1958  0.7754  -0.2525  0.8007
## depgroupcurrent:negE.pm    -1.5821  0.8749  -1.8083  0.0706

9.4.1.3 coefficients for the zero-inflated part

##                           Estimate Std.Err  z-value p-value
## (Intercept)                 0.7767  0.3904   1.9897  0.0466
## negE.pmc                   -1.5671  0.1398 -11.2079  0.0000
## negA_lag1.pmc              -1.3809  0.0691 -19.9836  0.0000
## depgroupremitted           -1.7747  0.4915  -3.6108  0.0003
## depgroupcurrent            -5.2901  0.6542  -8.0860  0.0000
## negE.pm                    -5.0347  2.9727  -1.6937  0.0903
## negE.pmc:depgroupremitted   0.0241  0.1767   0.1364  0.8915
## negE.pmc:depgroupcurrent    0.4184  0.3040   1.3763  0.1687
## depgroupremitted:negE.pm    0.4224  3.3999   0.1242  0.9011
## depgroupcurrent:negE.pm     6.4276  3.9139   1.6423  0.1005

9.4.1.4 exponential of coefficients for the zero-inflated part

##               (Intercept)                  negE.pmc             negA_lag1.pmc 
##                    2.1744                    0.2087                    0.2514 
##          depgroupremitted           depgroupcurrent                   negE.pm 
##                    0.1695                    0.0050                    0.0065 
## negE.pmc:depgroupremitted  negE.pmc:depgroupcurrent  depgroupremitted:negE.pm 
##                    1.0244                    1.5196                    1.5257 
##   depgroupcurrent:negE.pm 
##                  618.7106

9.4.1.5 random effects (intercept and slope (co)variance)

## $D
##                (Intercept)  negE.pmc negA_lag1.pmc zi_(Intercept)
## (Intercept)        0.39007 -0.041842     -0.050132       -1.34893
## negE.pmc          -0.04184  0.061003      0.004779        0.14951
## negA_lag1.pmc     -0.05013  0.004779      0.029016        0.07761
## zi_(Intercept)    -1.34893  0.149514      0.077612        6.87190
## attr(,"L")
##  [1]  0.624555 -0.066995 -0.080268 -2.159830  0.237729 -0.002519  0.020264
##  [8]  0.150223 -0.637065  1.341928

9.4.1.6 DHARMa residual plots

DHARMa_simres <- simulateResiduals(fittedModel = NA_negE_twopart_4)
par(mfrow=c(1,2))
plotQQunif(DHARMa_simres, testUniformity = F, testOutliers = F, testDispersion = F)
plotResiduals(DHARMa_simres, rank = T, smoothScatter = F)

9.4.1.7 likelihood ratio test

anova(NA_negE_twopart_3, NA_negE_twopart_4)

## 
##                        AIC      BIC   log.Lik     LRT df p.value
## NA_negE_twopart_3 39951.26 40055.11 -19948.63                   
## NA_negE_twopart_4 33194.98 33314.22 -16566.49 6764.28  4 <0.0001

9.4.2 Positive events

NA_posE_twopart_4 <- mixed_model(fixed=negA ~  posE.pmc + negA_lag1.pmc + depgroup + posE.pm + depgroup*posE.pmc + depgroup*posE.pm, random=~1 + posE.pmc + negA_lag1.pmc |pident,
                               data = emaD,
                               zi_fixed = ~ posE.pmc + negA_lag1.pmc + depgroup + posE.pm + depgroup*posE.pmc + depgroup*posE.pm,
                               zi_random= ~ 1|pident,
                               family = hurdle.lognormal(), n_phis=1, control=(iter_EM = 0))

9.4.2.1 model overview

## $logLik
## 'log Lik.' -16885.42 (df=31)
## 
## $AIC
## [1] 33832.84
## 
## $BIC
## [1] 33952.08
## 
## $call
## mixed_model(fixed = negA ~ posE.pmc + negA_lag1.pmc + depgroup + 
##     posE.pm + depgroup * posE.pmc + depgroup * posE.pm, random = ~1 + 
##     posE.pmc + negA_lag1.pmc | pident, data = emaD, family = hurdle.lognormal(), 
##     zi_fixed = ~posE.pmc + negA_lag1.pmc + depgroup + posE.pm + 
##         depgroup * posE.pmc + depgroup * posE.pm, zi_random = ~1 | 
##         pident, n_phis = 1, control = (iter_EM = 0))
## 
## $N
## [1] 16845
## 
## $phis_table
##        Estimate     Std.Err
## phi_1 -0.508605 0.007140645
## 
## $family
## 
## Family: two-part log-normal 
## Link function: identity

9.4.2.2 coefficients for the continuous part

##                           Estimate Std.Err z-value p-value
## (Intercept)                -1.0441  0.1087 -9.6009  0.0000
## posE.pmc                   -0.0838  0.0447 -1.8737  0.0610
## negA_lag1.pmc               0.3591  0.0172 20.8995  0.0000
## depgroupremitted            0.4340  0.1355  3.2031  0.0014
## depgroupcurrent             1.1489  0.1654  6.9469  0.0000
## posE.pm                    -0.4437  0.2558 -1.7347  0.0828
## posE.pmc:depgroupremitted  -0.1048  0.0497 -2.1110  0.0348
## posE.pmc:depgroupcurrent   -0.1006  0.0593 -1.6953  0.0900
## depgroupremitted:posE.pm    0.1327  0.3240  0.4095  0.6822
## depgroupcurrent:posE.pm    -0.1260  0.4355 -0.2892  0.7724

9.4.2.3 coefficients for the zero-inflated part

##                           Estimate Std.Err  z-value p-value
## (Intercept)                -0.0589  0.4607  -0.1279  0.8982
## posE.pmc                    0.7913  0.0991   7.9853  0.0000
## negA_lag1.pmc              -1.3551  0.0686 -19.7651  0.0000
## depgroupremitted           -1.7665  0.5886  -3.0015  0.0027
## depgroupcurrent            -4.9101  0.7773  -6.3167  0.0000
## posE.pm                     0.9349  1.0770   0.8680  0.3854
## posE.pmc:depgroupremitted  -0.1073  0.1216  -0.8830  0.3772
## posE.pmc:depgroupcurrent   -0.1438  0.1971  -0.7294  0.4658
## depgroupremitted:posE.pm   -0.3205  1.4039  -0.2283  0.8194
## depgroupcurrent:posE.pm     1.2852  2.0305   0.6330  0.5268

9.4.2.4 exponential of coefficients for the zero-inflated part

##               (Intercept)                  posE.pmc             negA_lag1.pmc 
##                    0.9428                    2.2063                    0.2579 
##          depgroupremitted           depgroupcurrent                   posE.pm 
##                    0.1709                    0.0074                    2.5468 
## posE.pmc:depgroupremitted  posE.pmc:depgroupcurrent  depgroupremitted:posE.pm 
##                    0.8982                    0.8661                    0.7258 
##   depgroupcurrent:posE.pm 
##                    3.6155

9.4.2.5 random effects (intercept SD and variance)

## $D
##                (Intercept)  posE.pmc negA_lag1.pmc zi_(Intercept)
## (Intercept)        0.37449 -0.026318     -0.043154       -1.30702
## posE.pmc          -0.02632  0.034081     -0.005129        0.11047
## negA_lag1.pmc     -0.04315 -0.005129      0.026389        0.03966
## zi_(Intercept)    -1.30702  0.110468      0.039659        6.92841
## attr(,"L")
##  [1]  0.61196 -0.04301 -0.07052 -2.13581  0.17953 -0.04546  0.10367  0.13910
##  [9] -0.76377  1.33140

9.4.2.6 DHARMa residual plots

DHARMa_simres <- simulateResiduals(fittedModel = NA_posE_twopart_4)
par(mfrow=c(1,2))
plotQQunif(DHARMa_simres, testUniformity = F, testOutliers = F, testDispersion = F)
plotResiduals(DHARMa_simres, rank = T, smoothScatter = F)

9.4.2.7 likelihood ratio test

anova(NA_posE_twopart_3, NA_posE_twopart_4)

## 
##                        AIC      BIC   log.Lik     LRT df p.value
## NA_posE_twopart_3 40783.86 40887.72 -20364.93                   
## NA_posE_twopart_4 33832.84 33952.08 -16885.42 6959.02  4 <0.0001

9.5 Final two-part models with current group as reference group

9.5.1 Negative events

NA_negE_twopart_4.2 <- mixed_model(fixed=negA ~  negE.pmc + negA_lag1.pmc + depgroup2 + negE.pm + depgroup2*negE.pmc + depgroup2*negE.pm, random=~1 + negE.pmc + negA_lag1.pmc|pident,
                               data = emaD,
                               zi_fixed = ~ negE.pmc + negA_lag1.pmc + depgroup2 + negE.pm + depgroup2*negE.pmc + depgroup2*negE.pm,
                               zi_random= ~ 1|pident,
                               family = hurdle.lognormal(), n_phis=1, control=(iter_EM = 0))

9.5.1.1 model overview

## $logLik
## 'log Lik.' -16567.54 (df=31)
## 
## $AIC
## [1] 33197.08
## 
## $BIC
## [1] 33316.32
## 
## $call
## mixed_model(fixed = negA ~ negE.pmc + negA_lag1.pmc + depgroup2 + 
##     negE.pm + depgroup2 * negE.pmc + depgroup2 * negE.pm, random = ~1 + 
##     negE.pmc + negA_lag1.pmc | pident, data = emaD, family = hurdle.lognormal(), 
##     zi_fixed = ~negE.pmc + negA_lag1.pmc + depgroup2 + negE.pm + 
##         depgroup2 * negE.pmc + depgroup2 * negE.pm, zi_random = ~1 | 
##         pident, n_phis = 1, control = (iter_EM = 0))
## 
## $N
## [1] 16845
## 
## $phis_table
##         Estimate     Std.Err
## phi_1 -0.5303653 0.007147631
## 
## $family
## 
## Family: two-part log-normal 
## Link function: identity

9.5.1.2 coefficients for the continuous part

##                            Estimate Std.Err z-value p-value
## (Intercept)                 -0.0216  0.1086 -0.1994  0.8420
## negE.pmc                     0.4498  0.0514  8.7524  0.0000
## negA_lag1.pmc                0.3745  0.0177 21.1883  0.0000
## depgroup2remitted           -0.8006  0.1266 -6.3252  0.0000
## depgroup2control            -1.3039  0.1426 -9.1460  0.0000
## negE.pm                     -0.7531  0.5362 -1.4046  0.1601
## negE.pmc:depgroup2remitted  -0.0033  0.0595 -0.0560  0.9553
## negE.pmc:depgroup2control   -0.0263  0.0730 -0.3599  0.7190
## depgroup2remitted:negE.pm    1.3743  0.6379  2.1544  0.0312
## depgroup2control:negE.pm     1.5715  0.8699  1.8065  0.0708

9.5.1.3 coefficients for the zero-inflated part

##                            Estimate Std.Err  z-value p-value
## (Intercept)                 -4.0080  0.5119  -7.8294  0.0000
## negE.pmc                    -1.1230  0.2678  -4.1937  0.0000
## negA_lag1.pmc               -1.3744  0.0689 -19.9578  0.0000
## depgroup2remitted            3.0334  0.5850   5.1854  0.0000
## depgroup2control             4.8819  0.6392   7.6379  0.0000
## negE.pm                      1.1235  2.4740   0.4541  0.6498
## negE.pmc:depgroup2remitted  -0.4221  0.2887  -1.4619  0.1438
## negE.pmc:depgroup2control   -0.4420  0.3020  -1.4638  0.1432
## depgroup2remitted:negE.pm   -5.8251  2.9191  -1.9955  0.0460
## depgroup2control:negE.pm    -6.3151  3.8410  -1.6441  0.1001

9.5.1.4 exponential of coefficients for the zero-inflated part

##                (Intercept)                   negE.pmc 
##                     0.0182                     0.3253 
##              negA_lag1.pmc          depgroup2remitted 
##                     0.2530                    20.7675 
##           depgroup2control                    negE.pm 
##                   131.8859                     3.0755 
## negE.pmc:depgroup2remitted  negE.pmc:depgroup2control 
##                     0.6557                     0.6427 
##  depgroup2remitted:negE.pm   depgroup2control:negE.pm 
##                     0.0030                     0.0018

9.5.1.5 random effects (intercept and slope (co)variance)

## $D
##                (Intercept)  negE.pmc negA_lag1.pmc zi_(Intercept)
## (Intercept)        0.38711 -0.041707     -0.050712       -1.29886
## negE.pmc          -0.04171  0.061406      0.005555        0.12372
## negA_lag1.pmc     -0.05071  0.005555      0.028157        0.07061
## zi_(Intercept)    -1.29886  0.123719      0.070613        6.57204
## attr(,"L")
##  [1]  0.6221785 -0.0670342 -0.0815066 -2.0876014  0.2385630  0.0003817
##  [7] -0.0679958  0.1466755 -0.6784663  1.3225062

9.5.2 Positive events

NA_posE_twopart_4.2 <- mixed_model(fixed=negA ~  posE.pmc + negA_lag1.pmc + depgroup2 + posE.pm + depgroup2*posE.pmc + depgroup2*posE.pm, random=~1 + posE.pmc + negA_lag1.pmc |pident,
                               data = emaD,
                               zi_fixed = ~ posE.pmc + negA_lag1.pmc + depgroup2 + posE.pm + depgroup2*posE.pmc + depgroup2*posE.pm,
                               zi_random= ~ 1|pident,
                               family = hurdle.lognormal(), n_phis=1, control=(iter_EM = 0))

9.5.2.1 model overview

## $logLik
## 'log Lik.' -16885.5 (df=31)
## 
## $AIC
## [1] 33832.99
## 
## $BIC
## [1] 33952.23
## 
## $call
## mixed_model(fixed = negA ~ posE.pmc + negA_lag1.pmc + depgroup2 + 
##     posE.pm + depgroup2 * posE.pmc + depgroup2 * posE.pm, random = ~1 + 
##     posE.pmc + negA_lag1.pmc | pident, data = emaD, family = hurdle.lognormal(), 
##     zi_fixed = ~posE.pmc + negA_lag1.pmc + depgroup2 + posE.pm + 
##         depgroup2 * posE.pmc + depgroup2 * posE.pm, zi_random = ~1 | 
##         pident, n_phis = 1, control = (iter_EM = 0))
## 
## $N
## [1] 16845
## 
## $phis_table
##         Estimate     Std.Err
## phi_1 -0.5086112 0.007141146
## 
## $family
## 
## Family: two-part log-normal 
## Link function: identity

9.5.2.2 coefficients for the continuous part

##                            Estimate Std.Err z-value p-value
## (Intercept)                  0.0823  0.1259  0.6535  0.5134
## posE.pmc                    -0.1832  0.0395 -4.6423  0.0000
## negA_lag1.pmc                0.3592  0.0173 20.7397  0.0000
## depgroup2remitted           -0.6869  0.1494 -4.5975  0.0000
## depgroup2control            -1.1407  0.1658 -6.8793  0.0000
## posE.pm                     -0.5586  0.3540 -1.5780  0.1146
## posE.pmc:depgroup2remitted  -0.0057  0.0465 -0.1225  0.9025
## posE.pmc:depgroup2control    0.0987  0.0591  1.6700  0.0949
## depgroup2remitted:posE.pm    0.2345  0.4067  0.5766  0.5642
## depgroup2control:posE.pm     0.1332  0.4366  0.3051  0.7603

9.5.2.3 coefficients for the zero-inflated part

##                            Estimate Std.Err  z-value p-value
## (Intercept)                 -4.8028  0.6284  -7.6432  0.0000
## posE.pmc                     0.6476  0.1702   3.8059  0.0001
## negA_lag1.pmc               -1.3544  0.0685 -19.7631  0.0000
## depgroup2remitted            2.9498  0.7252   4.0675  0.0000
## depgroup2control             4.8292  0.7782   6.2054  0.0000
## posE.pm                      2.1105  1.7260   1.2228  0.2214
## posE.pmc:depgroup2remitted   0.0335  0.1842   0.1820  0.8556
## posE.pmc:depgroup2control    0.1571  0.1969   0.7976  0.4251
## depgroup2remitted:posE.pm   -1.4441  1.9488  -0.7410  0.4587
## depgroup2control:posE.pm    -1.2708  2.0351  -0.6244  0.5323

9.5.2.4 exponential of coefficients for the zero-inflated part

##                (Intercept)                   posE.pmc 
##                     0.0082                     1.9110 
##              negA_lag1.pmc          depgroup2remitted 
##                     0.2581                    19.1025 
##           depgroup2control                    posE.pm 
##                   125.1106                     8.2523 
## posE.pmc:depgroup2remitted  posE.pmc:depgroup2control 
##                     1.0341                     1.1701 
##  depgroup2remitted:posE.pm   depgroup2control:posE.pm 
##                     0.2360                     0.2806

9.5.2.5 random effects (intercept SD and variance)

## $D
##                (Intercept)  posE.pmc negA_lag1.pmc zi_(Intercept)
## (Intercept)        0.37696 -0.026577     -0.043567       -1.31759
## posE.pmc          -0.02658  0.033652     -0.004913        0.10956
## negA_lag1.pmc     -0.04357 -0.004913      0.026815        0.03747
## zi_(Intercept)    -1.31759  0.109555      0.037473        6.97733
## attr(,"L")
##  [1]  0.61397 -0.04329 -0.07096 -2.14602  0.17827 -0.04479  0.09346  0.14062
##  [9] -0.78667  1.32074

10 Analyses of transformed positive affect

emaD$posA_pow2 <- emaD$posA^2

10.1 negative events

Model_PApow_negE <- lmer(posA_pow2 ~ negE.pmc + posA_lag1.pmc + depgroup + negE.pm + depgroup*negE.pmc + depgroup*negE.pm + (negE.pmc + posA_lag1.pmc|pident), data = emaD)

summary(Model_PApow_negE)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: 
## posA_pow2 ~ negE.pmc + posA_lag1.pmc + depgroup + negE.pm + depgroup *  
##     negE.pmc + depgroup * negE.pm + (negE.pmc + posA_lag1.pmc |      pident)
##    Data: emaD
## 
## REML criterion at convergence: 98900.4
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -5.8208 -0.5716  0.0150  0.5629  4.7907 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr       
##  pident   (Intercept)   26.0586  5.1048              
##           negE.pmc       4.5550  2.1343   -0.15      
##           posA_lag1.pmc  0.8008  0.8949    0.34  0.04
##  Residual               18.5727  4.3096              
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##                            Estimate Std. Error        df t value Pr(>|t|)    
## (Intercept)                20.71502    0.74351 348.72658  27.861  < 2e-16 ***
## negE.pmc                   -3.78323    0.34450 291.40790 -10.982  < 2e-16 ***
## posA_lag1.pmc               2.31609    0.07133 310.46138  32.470  < 2e-16 ***
## depgroupremitted           -4.32453    0.92614 344.70675  -4.669 4.33e-06 ***
## depgroupcurrent           -11.31879    1.18589 338.35289  -9.545  < 2e-16 ***
## negE.pm                   -12.71440    5.60459 341.91889  -2.269   0.0239 *  
## negE.pmc:depgroupremitted   0.28369    0.41153 275.60303   0.689   0.4912    
## negE.pmc:depgroupcurrent    0.47267    0.52604 274.24491   0.899   0.3697    
## depgroupremitted:negE.pm    4.32114    6.34502 340.03905   0.681   0.4963    
## depgroupcurrent:negE.pm    18.07862    7.25174 335.85412   2.493   0.0131 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##                (Intr) ngE.pmc psA_1. dpgrpr dpgrpc negE.pm ngE.pmc:dpgrpr
## negE.pmc       -0.039                                                    
## psA_lg1.pmc     0.062  0.020                                             
## depgrprmttd    -0.799  0.033   0.017                                     
## depgrpcrrnt    -0.623  0.026   0.019  0.502                              
## negE.pm        -0.719 -0.042   0.009  0.578  0.452                       
## ngE.pmc:dpgrpr  0.034 -0.837  -0.002 -0.042 -0.021  0.035                
## ngE.pmc:dpgrpc  0.027 -0.655  -0.002 -0.021 -0.054  0.028   0.548        
## dpgrprmt:E.     0.636  0.037  -0.003 -0.722 -0.399 -0.883  -0.041        
## dpgrpcrr:E.     0.556  0.032  -0.003 -0.447 -0.707 -0.773  -0.027        
##                ngE.pmc:dpgrpc dpgrpr:E.
## negE.pmc                               
## psA_lg1.pmc                            
## depgrprmttd                            
## depgrpcrrnt                            
## negE.pm                                
## ngE.pmc:dpgrpr                         
## ngE.pmc:dpgrpc                         
## dpgrprmt:E.    -0.024                  
## dpgrpcrr:E.    -0.029          0.683

10.1.0.1 BIC

## [1] 99065.86

10.1.1 DHARMa residual plots

DHARMa_simres <- simulateResiduals(fittedModel = Model_PApow_negE)
par(mfrow=c(1,2))
plotQQunif(DHARMa_simres, testUniformity = F, testOutliers = F, testDispersion = F)
plotResiduals(DHARMa_simres, rank = T, smoothScatter = F)

10.2 positive events

Model_PApow_posE <- lmer(posA_pow2 ~ posE.pmc + posA_lag1.pmc + depgroup + posE.pm + depgroup*posE.pmc + depgroup*posE.pm + (posE.pmc + posA_lag1.pmc|pident), data = emaD)

summary(Model_PApow_posE)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: 
## posA_pow2 ~ posE.pmc + posA_lag1.pmc + depgroup + posE.pm + depgroup *  
##     posE.pmc + depgroup * posE.pm + (posE.pmc + posA_lag1.pmc |      pident)
##    Data: emaD
## 
## REML criterion at convergence: 98975.6
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -5.6643 -0.5535  0.0352  0.5785  4.6385 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr       
##  pident   (Intercept)   26.8317  5.1799              
##           posE.pmc       2.2991  1.5163   -0.15      
##           posA_lag1.pmc  0.8845  0.9405    0.34 -0.21
##  Residual               18.6307  4.3163              
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##                            Estimate Std. Error        df t value Pr(>|t|)    
## (Intercept)                19.45942    0.88371 347.08103  22.020  < 2e-16 ***
## posE.pmc                    2.03502    0.22827 322.59488   8.915  < 2e-16 ***
## posA_lag1.pmc               2.21395    0.07342 309.37472  30.154  < 2e-16 ***
## depgroupremitted           -4.99680    1.10938 342.24990  -4.504 9.16e-06 ***
## depgroupcurrent           -11.00708    1.40412 338.44444  -7.839 5.96e-14 ***
## posE.pm                    -0.03351    2.05892 345.24261  -0.016   0.9870    
## posE.pmc:depgroupremitted   0.56038    0.27652 311.45999   2.027   0.0436 *  
## posE.pmc:depgroupcurrent    0.53757    0.36307 318.31522   1.481   0.1397    
## depgroupremitted:posE.pm    2.16527    2.62747 341.76891   0.824   0.4105    
## depgroupcurrent:posE.pm     6.26503    3.71999 333.32344   1.684   0.0931 .  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##                (Intr) psE.pmc psA_1. dpgrpr dpgrpc posE.pm psE.pmc:dpgrpr
## posE.pmc       -0.042                                                    
## psA_lg1.pmc     0.062 -0.068                                             
## depgrprmttd    -0.793  0.029   0.011                                     
## depgrpcrrnt    -0.626  0.023   0.009  0.500                              
## posE.pm        -0.806 -0.011  -0.002  0.642  0.507                       
## psE.pmc:dpgrpr  0.030 -0.821  -0.009 -0.037 -0.020  0.009                
## psE.pmc:dpgrpc  0.022 -0.624  -0.020 -0.019 -0.036  0.007   0.517        
## dpgrprmt:E.     0.631  0.008   0.002 -0.818 -0.397 -0.784  -0.010        
## dpgrpcrr:E.     0.446  0.005   0.009 -0.355 -0.803 -0.553  -0.005        
##                psE.pmc:dpgrpc dpgrpr:E.
## posE.pmc                               
## psA_lg1.pmc                            
## depgrprmttd                            
## depgrpcrrnt                            
## posE.pm                                
## psE.pmc:dpgrpr                         
## psE.pmc:dpgrpc                         
## dpgrprmt:E.    -0.005                  
## dpgrpcrr:E.    -0.008          0.434

10.2.0.1 BIC

## [1] 99141.06

10.2.1 DHARMa residual plots

DHARMa_simres <- simulateResiduals(fittedModel = Model_PApow_posE)
par(mfrow=c(1,2))
plotQQunif(DHARMa_simres, testUniformity = F, testOutliers = F, testDispersion = F)
plotResiduals(DHARMa_simres, rank = T, smoothScatter = F)

11 Analysis of PA-reactivity to positive events including momentart event * event load interaction

Model_PA_posE_int <- lmer(posA ~ posE.pmc + posA_lag1.pmc + depgroup + posE.pm + posE.pmc*posE.pm + depgroup*posE.pmc + depgroup*posE.pm + (posE.pmc + posA_lag1.pmc|pident), data = emaD)

summary(Model_PA_posE_int)

## Linear mixed model fit by REML. t-tests use Satterthwaite's method [
## lmerModLmerTest]
## Formula: posA ~ posE.pmc + posA_lag1.pmc + depgroup + posE.pm + posE.pmc *  
##     posE.pm + depgroup * posE.pmc + depgroup * posE.pm + (posE.pmc +  
##     posA_lag1.pmc | pident)
##    Data: emaD
## 
## REML criterion at convergence: 33839
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -6.0687 -0.4795  0.0927  0.5770  3.8208 
## 
## Random effects:
##  Groups   Name          Variance Std.Dev. Corr       
##  pident   (Intercept)   0.48416  0.6958              
##           posE.pmc      0.04904  0.2215   -0.44      
##           posA_lag1.pmc 0.01906  0.1381   -0.05 -0.14
##  Residual               0.39013  0.6246              
## Number of obs: 16845, groups:  pident, 346
## 
## Fixed effects:
##                            Estimate Std. Error        df t value Pr(>|t|)    
## (Intercept)                 4.33538    0.11943 356.08898  36.301  < 2e-16 ***
## posE.pmc                    0.16503    0.04537 386.07813   3.637 0.000313 ***
## posA_lag1.pmc               0.31461    0.01074 343.01933  29.301  < 2e-16 ***
## depgroupremitted           -0.67544    0.14908 358.24198  -4.531 8.02e-06 ***
## depgroupcurrent            -1.57544    0.18942 358.35632  -8.317 1.90e-15 ***
## posE.pm                    -0.01255    0.27615 357.97578  -0.045 0.963775    
## posE.pmc:posE.pm            0.20320    0.08205 401.65999   2.476 0.013680 *  
## posE.pmc:depgroupremitted   0.12571    0.04023 318.82279   3.124 0.001946 ** 
## posE.pmc:depgroupcurrent    0.20301    0.05319 325.93909   3.817 0.000162 ***
## depgroupremitted:posE.pm    0.27331    0.34834 351.41300   0.785 0.433219    
## depgroupcurrent:posE.pm     0.99667    0.49618 351.93253   2.009 0.045334 *  
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##                (Intr) psE.pmc psA_1. dpgrpr dpgrpc posE.pm pE.:E.
## posE.pmc       -0.196                                            
## psA_lg1.pmc    -0.008 -0.033                                     
## depgrprmttd    -0.787  0.081  -0.001                             
## depgrpcrrnt    -0.622  0.078  -0.001  0.497                      
## posE.pm        -0.800  0.079   0.001  0.623  0.494               
## psE.pmc:pE.     0.128 -0.682  -0.010  0.009 -0.013 -0.161        
## psE.pmc:dpgrpr  0.124 -0.604  -0.005 -0.147 -0.078  0.034   0.006
## psE.pmc:dpgrpc  0.108 -0.530  -0.017 -0.074 -0.158  0.008   0.112
## dpgrprmt:E.     0.617  0.032  -0.001 -0.808 -0.390 -0.770  -0.012
## dpgrpcrr:E.     0.434  0.018  -0.001 -0.348 -0.791 -0.542  -0.002
##                psE.pmc:dpgrpr psE.pmc:dpgrpc dpgrpr:E.
## posE.pmc                                              
## psA_lg1.pmc                                           
## depgrprmttd                                           
## depgrpcrrnt                                           
## posE.pm                                               
## psE.pmc:pE.                                           
## psE.pmc:dpgrpr                                        
## psE.pmc:dpgrpc  0.513                                 
## dpgrprmt:E.    -0.044         -0.022                  
## dpgrpcrr:E.    -0.019         -0.037          0.430

11.0.1 DHARMa residual plots

DHARMa_simres <- simulateResiduals(fittedModel = Model_PA_posE_int)
par(mfrow=c(1,2))
plotQQunif(DHARMa_simres, testUniformity = F, testOutliers = F, testDispersion = F)
plotResiduals(DHARMa_simres, rank = T, smoothScatter = F)

12 R session info

## R version 4.0.3 (2020-10-10)
## Platform: x86_64-apple-darwin17.0 (64-bit)
## Running under: macOS Catalina 10.15.7
## 
## Matrix products: default
## BLAS:   /Library/Frameworks/R.framework/Versions/4.0/Resources/lib/libRblas.dylib
## LAPACK: /Library/Frameworks/R.framework/Versions/4.0/Resources/lib/libRlapack.dylib
## 
## locale:
## [1] en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-8
## 
## attached base packages:
## [1] stats     graphics  grDevices utils     datasets  methods   base     
## 
## other attached packages:
##  [1] GLMMadaptive_0.8-0 DHARMa_0.4.3       effects_4.2-0      car_3.0-10        
##  [5] carData_3.0-4      stringr_1.4.0      reshape2_1.4.4     lmerTest_3.1-3    
##  [9] lme4_1.1-25        Matrix_1.2-18      FSA_0.8.30         ggpubr_0.4.0      
## [13] dplyr_1.0.2        knitr_1.33         moments_0.14       ggplot2_3.3.2     
## [17] plyr_1.8.6        
## 
## loaded via a namespace (and not attached):
##  [1] tidyr_1.1.2         splines_4.0.3       foreach_1.5.1      
##  [4] gap_1.2.2           statmod_1.4.35      highr_0.8          
##  [7] cellranger_1.1.0    yaml_2.2.1          numDeriv_2016.8-1.1
## [10] pillar_1.4.6        backports_1.2.0     lattice_0.20-41    
## [13] glue_1.4.2          digest_0.6.27       ggsignif_0.6.0     
## [16] minqa_1.2.4         colorspace_1.4-1    htmltools_0.5.0    
## [19] survey_4.0          pkgconfig_2.0.3     broom_0.7.2        
## [22] haven_2.3.1         purrr_0.3.4         scales_1.1.1       
## [25] openxlsx_4.2.3      rio_0.5.16          tibble_3.0.4       
## [28] generics_0.1.0      ellipsis_0.3.1      withr_2.3.0        
## [31] nnet_7.3-14         survival_3.2-7      magrittr_1.5       
## [34] crayon_1.3.4        readxl_1.3.1        evaluate_0.14      
## [37] nlme_3.1-149        MASS_7.3-53         rstatix_0.6.0      
## [40] forcats_0.5.0       foreign_0.8-80      tools_4.0.3        
## [43] data.table_1.13.2   hms_0.5.3           mitools_2.4        
## [46] matrixStats_0.57.0  lifecycle_0.2.0     munsell_0.5.0      
## [49] zip_2.1.1           compiler_4.0.3      rlang_0.4.8        
## [52] grid_4.0.3          nloptr_1.2.2.2      iterators_1.0.13   
## [55] rmarkdown_2.10      boot_1.3-25         gtable_0.3.0       
## [58] codetools_0.2-16    abind_1.4-5         DBI_1.1.0          
## [61] curl_4.3            R6_2.5.0            insight_0.10.0     
## [64] stringi_1.5.3       parallel_4.0.3      Rcpp_1.0.5         
## [67] vctrs_0.3.4         tidyselect_1.1.0    xfun_0.25